gitlab-org--gitlab-foss/doc/administration/reference_architectures/5k_users.md

stage	group	info
Enablement	Distribution	To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/engineering/ux/technical-writing/#assignments

Reference architecture: up to 5,000 users (PREMIUM SELF)

This page describes GitLab reference architecture for up to 5,000 users. For a full list of reference architectures, see Available reference architectures.

NOTE: This reference architecture is designed to help your organization achieve a highly-available GitLab deployment. If you do not have the expertise or need to maintain a highly-available environment, you can have a simpler and less costly-to-operate environment by using the 2,000-user reference architecture.

• Supported users (approximate): 5,000
• High Availability: Yes (Praefect needs a third-party PostgreSQL solution for HA)
• Cloud Native Hybrid: Yes
• Test requests per second (RPS) rates: API: 100 RPS, Web: 10 RPS, Git (Pull): 10 RPS, Git (Push): 2 RPS
• Latest 5k weekly performance testing results

Service	Nodes	Configuration	GCP	AWS	Azure
External load balancing node3	1	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Redis2	3	2 vCPU, 7.5 GB memory	n1-standard-2	m5.large	D2s v3
Consul1 + Sentinel2	3	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
PostgreSQL1	3	4 vCPU, 15 GB memory	n1-standard-4	m5.xlarge	D4s v3
PgBouncer1	3	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Internal load balancing node3	1	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Gitaly5	3	8 vCPU, 30 GB memory	n1-standard-8	m5.2xlarge	D8s v3
Praefect5	3	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Praefect PostgreSQL1	1+	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Sidekiq	4	2 vCPU, 7.5 GB memory	n1-standard-2	m5.large	D2s v3
GitLab Rails	3	16 vCPU, 14.4 GB memory	n1-highcpu-16	c5.4xlarge	F16s v2
Monitoring node	1	2 vCPU, 1.8 GB memory	n1-highcpu-2	c5.large	F2s v2
Object storage4	n/a	n/a	n/a	n/a	n/a
NFS server (optional, not recommended)	1	4 vCPU, 3.6 GB memory	n1-highcpu-4	c5.xlarge	F4s v2

1. Can be optionally run on reputable third-party external PaaS PostgreSQL solutions. Google Cloud SQL and AWS RDS are known to work, however Azure Database for PostgreSQL is not recommended due to performance issues. Consul is primarily used for PostgreSQL high availability so can be ignored when using a PostgreSQL PaaS setup. However it is also used optionally by Prometheus for Omnibus auto host discovery.
2. Can be optionally run on reputable third-party external PaaS Redis solutions. Google Memorystore and AWS Elasticache are known to work.
3. Can be optionally run on reputable third-party load balancing services (LB PaaS). AWS ELB is known to work.
4. Should be run on reputable third-party object storage (storage PaaS) for cloud implementations. Google Cloud Storage and AWS S3 are known to work.
5. Gitaly Cluster provides the benefits of fault tolerance, but comes with additional complexity of setup and management. Please review the existing technical limitations and considerations prior to deploying Gitaly Cluster. If Gitaly Sharded is desired, the same specs listed above for Gitaly should be used.

NOTE: For all PaaS solutions that involve configuring instances, it is strongly recommended to implement a minimum of three nodes in three different availability zones to align with resilient cloud architecture practices.

@startuml 5k
card "**External Load Balancer**" as elb #6a9be7
card "**Internal Load Balancer**" as ilb #9370DB

together {
  collections "**GitLab Rails** x3" as gitlab #32CD32
  collections "**Sidekiq** x4" as sidekiq #ff8dd1
}

card "**Prometheus + Grafana**" as monitor #7FFFD4

card "Gitaly Cluster" as gitaly_cluster {
  collections "**Praefect** x3" as praefect #FF8C00
  collections "**Gitaly** x3" as gitaly #FF8C00
  card "**Praefect PostgreSQL***\n//Non fault-tolerant//" as praefect_postgres #FF8C00

  praefect -[#FF8C00]-> gitaly
  praefect -[#FF8C00]> praefect_postgres
}

card "Database" as database {
  collections "**PGBouncer** x3" as pgbouncer #4EA7FF
  card "**PostgreSQL** (Primary)" as postgres_primary #4EA7FF
  collections "**PostgreSQL** (Secondary) x2" as postgres_secondary #4EA7FF

  pgbouncer -[#4EA7FF]-> postgres_primary
  postgres_primary .[#4EA7FF]> postgres_secondary
}

card "**Consul + Sentinel**" as consul_sentinel {
 collections "**Consul** x3" as consul #e76a9b
 collections "**Redis Sentinel** x3" as sentinel #e6e727
}

card "Redis" as redis {
  collections "**Redis** x3" as redis_nodes #FF6347

  redis_nodes <.[#FF6347]- sentinel
}

cloud "**Object Storage**" as object_storage #white

elb -[#6a9be7]-> gitlab
elb -[#6a9be7]--> monitor

gitlab -[#32CD32]--> ilb
gitlab -[#32CD32]-> object_storage
gitlab -[#32CD32]---> redis
gitlab -[hidden]-> monitor
gitlab -[hidden]-> consul

sidekiq -[#ff8dd1]--> ilb
sidekiq -[#ff8dd1]-> object_storage
sidekiq -[#ff8dd1]---> redis
sidekiq -[hidden]-> monitor
sidekiq -[hidden]-> consul

ilb -[#9370DB]-> gitaly_cluster
ilb -[#9370DB]-> database

consul .[#e76a9b]u-> gitlab
consul .[#e76a9b]u-> sidekiq
consul .[#e76a9b]> monitor
consul .[#e76a9b]-> database
consul .[#e76a9b]-> gitaly_cluster
consul .[#e76a9b,norank]--> redis

monitor .[#7FFFD4]u-> gitlab
monitor .[#7FFFD4]u-> sidekiq
monitor .[#7FFFD4]> consul
monitor .[#7FFFD4]-> database
monitor .[#7FFFD4]-> gitaly_cluster
monitor .[#7FFFD4,norank]--> redis
monitor .[#7FFFD4]> ilb
monitor .[#7FFFD4,norank]u--> elb

@enduml

The Google Cloud Platform (GCP) architectures were built and tested using the Intel Xeon E5 v3 (Haswell) CPU platform. On different hardware you may find that adjustments, either lower or higher, are required for your CPU or node counts. For more information, see our Sysbench-based CPU benchmarks.

Due to better performance and availability, for data objects (such as LFS, uploads, or artifacts), using an object storage service is recommended instead of using NFS. Using an object storage service also doesn't require you to provision and maintain a node.

It's also worth noting that at this time Praefect requires its own database server and that to achieve full High Availability a third-party PostgreSQL database solution will be required. We hope to offer a built in solutions for these restrictions in the future but in the meantime a non HA PostgreSQL server can be set up via Omnibus GitLab, which the above specs reflect. Refer to the following issues for more information: omnibus-gitlab#5919 & gitaly#3398

Setup components

To set up GitLab and its components to accommodate up to 5,000 users:

1. Configure the external load balancer to handle the load balancing of the GitLab application services nodes.
2. Configure the internal load balancer. to handle the load balancing of GitLab application internal connections.
3. Configure Redis.
4. Configure Consul and Sentinel.
5. Configure PostgreSQL, the database for GitLab.
6. Configure PgBouncer.
7. Configure Gitaly Cluster, provides access to the Git repositories.
8. Configure Sidekiq.
9. Configure the main GitLab Rails application to run Puma, Workhorse, GitLab Shell, and to serve all frontend requests (which include UI, API, and Git over HTTP/SSH).
10. Configure Prometheus to monitor your GitLab environment.
11. Configure the object storage used for shared data objects.
12. Configure Advanced Search (optional) for faster, more advanced code search across your entire GitLab instance.
13. Configure NFS (optional, and not recommended) to have shared disk storage service as an alternative to Gitaly or object storage. You can skip this step if you're not using GitLab Pages (which requires NFS).

The servers start on the same 10.6.0.0/24 private network range, and can connect to each other freely on these addresses.

The following list includes descriptions of each server and its assigned IP:

• 10.6.0.10: External Load Balancer
• 10.6.0.61: Redis Primary
• 10.6.0.62: Redis Replica 1
• 10.6.0.63: Redis Replica 2
• 10.6.0.11: Consul/Sentinel 1
• 10.6.0.12: Consul/Sentinel 2
• 10.6.0.13: Consul/Sentinel 3
• 10.6.0.31: PostgreSQL primary
• 10.6.0.32: PostgreSQL secondary 1
• 10.6.0.33: PostgreSQL secondary 2
• 10.6.0.21: PgBouncer 1
• 10.6.0.22: PgBouncer 2
• 10.6.0.23: PgBouncer 3
• 10.6.0.20: Internal Load Balancer
• 10.6.0.51: Gitaly 1
• 10.6.0.52: Gitaly 2
• 10.6.0.93: Gitaly 3
• 10.6.0.131: Praefect 1
• 10.6.0.132: Praefect 2
• 10.6.0.133: Praefect 3
• 10.6.0.141: Praefect PostgreSQL 1 (non HA)
• 10.6.0.71: Sidekiq 1
• 10.6.0.72: Sidekiq 2
• 10.6.0.73: Sidekiq 3
• 10.6.0.74: Sidekiq 4
• 10.6.0.41: GitLab application 1
• 10.6.0.42: GitLab application 2
• 10.6.0.43: GitLab application 3
• 10.6.0.81: Prometheus

Configure the external load balancer

In a multi-node GitLab configuration, you'll need a load balancer to route traffic to the application servers. The specifics on which load balancer to use or its exact configuration is beyond the scope of GitLab documentation. We assume that if you're managing multi-node systems like GitLab, you already have a load balancer of choice and that the routing methods used are distributing calls evenly between all nodes. Some load balancer examples include HAProxy (open-source), F5 Big-IP LTM, and Citrix Net Scaler. This documentation outline the ports and protocols needed for use with GitLab.

This architecture has been tested and validated with HAProxy as the load balancer. Although other load balancers with similar feature sets could also be used, those load balancers have not been validated.

The next question is how you will handle SSL in your environment. There are several different options:

• The application node terminates SSL.
• The load balancer terminates SSL without backend SSL and communication is not secure between the load balancer and the application node.
• The load balancer terminates SSL with backend SSL and communication is secure between the load balancer and the application node.

Application node terminates SSL

Configure your load balancer to pass connections on port 443 as TCP rather than HTTP(S) protocol. This will pass the connection to the application node's NGINX service untouched. NGINX will have the SSL certificate and listen on port 443.

See the NGINX HTTPS documentation for details on managing SSL certificates and configuring NGINX.

Load balancer terminates SSL without backend SSL

Configure your load balancer to use the HTTP(S) protocol rather than TCP. The load balancer will then be responsible for managing SSL certificates and terminating SSL.

Since communication between the load balancer and GitLab will not be secure, there is some additional configuration needed. See the NGINX proxied SSL documentation for details.

Load balancer terminates SSL with backend SSL

Configure your load balancer(s) to use the 'HTTP(S)' protocol rather than 'TCP'. The load balancer(s) will be responsible for managing SSL certificates that end users will see.

Traffic will also be secure between the load balancer(s) and NGINX in this scenario. There is no need to add configuration for proxied SSL since the connection will be secure all the way. However, configuration will need to be added to GitLab to configure SSL certificates. See NGINX HTTPS documentation for details on managing SSL certificates and configuring NGINX.

Readiness checks

Ensure the external load balancer only routes to working services with built in monitoring endpoints. The readiness checks all require additional configuration on the nodes being checked, otherwise, the external load balancer will not be able to connect.

Ports

The basic ports to be used are shown in the table below.

LB Port	Backend Port	Protocol
80	80	HTTP (1)
443	443	TCP or HTTPS (1) (2)
22	22	TCP

• (1): Web terminal support requires your load balancer to correctly handle WebSocket connections. When using HTTP or HTTPS proxying, this means your load balancer must be configured to pass through the Connection and Upgrade hop-by-hop headers. See the web terminal integration guide for more details.
• (2): When using HTTPS protocol for port 443, you will need to add an SSL certificate to the load balancers. If you wish to terminate SSL at the GitLab application server instead, use TCP protocol.

If you're using GitLab Pages with custom domain support you will need some additional port configurations. GitLab Pages requires a separate virtual IP address. Configure DNS to point the pages_external_url from /etc/gitlab/gitlab.rb at the new virtual IP address. See the GitLab Pages documentation for more information.

LB Port	Backend Port	Protocol
80	Varies (1)	HTTP
443	Varies (1)	TCP (2)

• (1): The backend port for GitLab Pages depends on the gitlab_pages['external_http'] and gitlab_pages['external_https'] setting. See GitLab Pages documentation for more details.
• (2): Port 443 for GitLab Pages should always use the TCP protocol. Users can configure custom domains with custom SSL, which would not be possible if SSL was terminated at the load balancer.

Alternate SSH Port

Some organizations have policies against opening SSH port 22. In this case, it may be helpful to configure an alternate SSH hostname that allows users to use SSH on port 443. An alternate SSH hostname will require a new virtual IP address compared to the other GitLab HTTP configuration above.

Configure DNS for an alternate SSH hostname such as altssh.gitlab.example.com.

LB Port	Backend Port	Protocol
443	22	TCP

Back to setup components

Configure the internal load balancer

The Internal Load Balancer is used to balance any internal connections the GitLab environment requires such as connections to PgBouncer and Praefect (Gitaly Cluster).

It's a separate node from the External Load Balancer and shouldn't have any access externally.

The following IP will be used as an example:

• 10.6.0.40: Internal Load Balancer

Here's how you could do it with HAProxy:

global
    log /dev/log local0
    log localhost local1 notice
    log stdout format raw local0

defaults
    log global
    default-server inter 10s fall 3 rise 2
    balance leastconn

frontend internal-pgbouncer-tcp-in
    bind *:6432
    mode tcp
    option tcplog

    default_backend pgbouncer

frontend internal-praefect-tcp-in
    bind *:2305
    mode tcp
    option tcplog
    option clitcpka

    default_backend praefect

backend pgbouncer
    mode tcp
    option tcp-check

    server pgbouncer1 10.6.0.21:6432 check
    server pgbouncer2 10.6.0.22:6432 check
    server pgbouncer3 10.6.0.23:6432 check

backend praefect
    mode tcp
    option tcp-check
    option srvtcpka

    server praefect1 10.6.0.131:2305 check
    server praefect2 10.6.0.132:2305 check
    server praefect3 10.6.0.133:2305 check

Refer to your preferred Load Balancer's documentation for further guidance. Also ensure that the routing methods used are distributing calls evenly across all nodes.

Back to setup components

Configure Redis

Using Redis in scalable environment is possible using a Primary x Replica topology with a Redis Sentinel service to watch and automatically start the failover procedure.

Redis requires authentication if used with Sentinel. See Redis Security documentation for more information. We recommend using a combination of a Redis password and tight firewall rules to secure your Redis service. You are highly encouraged to read the Redis Sentinel documentation before configuring Redis with GitLab to fully understand the topology and architecture.

In this section, you'll be guided through configuring an external Redis instance to be used with GitLab. The following IPs will be used as an example:

• 10.6.0.61: Redis Primary
• 10.6.0.62: Redis Replica 1
• 10.6.0.63: Redis Replica 2

Provide your own Redis instance

Managed Redis from cloud providers such as AWS ElastiCache will work. If these services support high availability, be sure it is not the Redis Cluster type.

Redis version 5.0 or higher is required, as this is what ships with Omnibus GitLab packages starting with GitLab 13.0. Older Redis versions do not support an optional count argument to SPOP which is now required for Merge Trains.

Note the Redis node's IP address or hostname, port, and password (if required). These will be necessary when configuring the GitLab application servers later.

Standalone Redis using Omnibus GitLab

This is the section where we install and set up the new Redis instances.

The requirements for a Redis setup are the following:

1. All Redis nodes must be able to talk to each other and accept incoming connections over Redis (6379) and Sentinel (26379) ports (unless you change the default ones).
2. The server that hosts the GitLab application must be able to access the Redis nodes.
3. Protect the nodes from access from external networks (Internet), using a firewall.

Both the primary and replica Redis nodes need the same password defined in redis['password']. At any time during a failover, the Sentinels can reconfigure a node and change its status from primary to replica (and vice versa).

Configuring the primary Redis instance

1. SSH in to the Primary Redis server.

2. Download and install the Omnibus GitLab package of your choice. Be sure to both follow only installation steps 1 and 2 on the page, and to select the correct Omnibus GitLab package, with the same version and type (Community or Enterprise editions) as your current install.

3. Edit /etc/gitlab/gitlab.rb and add the contents:

   # Specify server role as 'redis_master_role' and enable Consul agent
   roles(['redis_master_role', 'consul_role'])
   
   # IP address pointing to a local IP that the other machines can reach to.
   # You can also set bind to '0.0.0.0' which listen in all interfaces.
   # If you really need to bind to an external accessible IP, make
   # sure you add extra firewall rules to prevent unauthorized access.
   redis['bind'] = '10.6.0.61'
   
   # Define a port so Redis can listen for TCP requests which will allow other
   # machines to connect to it.
   redis['port'] = 6379
   
   # Set up password authentication for Redis (use the same password in all nodes).
   redis['password'] = 'redis-password-goes-here'
   
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   
   # Set the network addresses that the exporters will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   redis_exporter['listen_address'] = '0.0.0.0:9121'
   redis_exporter['flags'] = {
        'redis.addr' => 'redis://10.6.0.61:6379',
        'redis.password' => 'redis-password-goes-here',
   }
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. Reconfigure Omnibus GitLab for the changes to take effect.

You can specify multiple roles, like sentinel and Redis, as: roles(['redis_sentinel_role', 'redis_master_role']). Read more about roles.

You can list the current Redis Primary, Replica status via:

/opt/gitlab/embedded/bin/redis-cli -h <host> -a 'redis-password-goes-here' info replication

Show running GitLab services via:

gitlab-ctl status

The output should be similar to the following:

run: consul: (pid 30043) 76863s; run: log: (pid 29691) 76892s
run: logrotate: (pid 31152) 3070s; run: log: (pid 29595) 76908s
run: node-exporter: (pid 30064) 76862s; run: log: (pid 29624) 76904s
run: redis: (pid 30070) 76861s; run: log: (pid 29573) 76914s
run: redis-exporter: (pid 30075) 76861s; run: log: (pid 29674) 76896s

Configuring the replica Redis instances

1. SSH in to the replica Redis server.

2. Download and install the Omnibus GitLab package of your choice. Be sure to both follow only installation steps 1 and 2 on the page, and to select the correct Omnibus GitLab package, with the same version and type (Community or Enterprise editions) as your current install.

3. Edit /etc/gitlab/gitlab.rb and add the contents:

   # Specify server role as 'redis_replica_role' and enable Consul agent
   roles(['redis_replica_role', 'consul_role'])
   
   # IP address pointing to a local IP that the other machines can reach to.
   # You can also set bind to '0.0.0.0' which listen in all interfaces.
   # If you really need to bind to an external accessible IP, make
   # sure you add extra firewall rules to prevent unauthorized access.
   redis['bind'] = '10.6.0.62'
   
   # Define a port so Redis can listen for TCP requests which will allow other
   # machines to connect to it.
   redis['port'] = 6379
   
   # The same password for Redis authentication you set up for the primary node.
   redis['password'] = 'redis-password-goes-here'
   
   # The IP of the primary Redis node.
   redis['master_ip'] = '10.6.0.61'
   
   # Port of primary Redis server, uncomment to change to non default. Defaults
   # to `6379`.
   #redis['master_port'] = 6379
   
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   
   # Set the network addresses that the exporters will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   redis_exporter['listen_address'] = '0.0.0.0:9121'
   redis_exporter['flags'] = {
        'redis.addr' => 'redis://10.6.0.62:6379',
        'redis.password' => 'redis-password-goes-here',
   }
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. Reconfigure Omnibus GitLab for the changes to take effect.

6. Go through the steps again for all the other replica nodes, and make sure to set up the IPs correctly.

You can specify multiple roles, like sentinel and Redis, as: roles(['redis_sentinel_role', 'redis_master_role']). Read more about roles.

These values don't have to be changed again in /etc/gitlab/gitlab.rb after a failover, as the nodes will be managed by the Sentinels, and even after a gitlab-ctl reconfigure, they will get their configuration restored by the same Sentinels.

Advanced configuration options are supported and can be added if needed.

Back to setup components

Configure Consul and Sentinel

Now that the Redis servers are all set up, let's configure the Sentinel servers. The following IPs will be used as an example:

• 10.6.0.11: Consul/Sentinel 1
• 10.6.0.12: Consul/Sentinel 2
• 10.6.0.13: Consul/Sentinel 3

NOTE: If you're using an external Redis Sentinel instance, be sure to exclude the requirepass parameter from the Sentinel configuration. This parameter causes clients to report NOAUTH Authentication required.. Redis Sentinel 3.2.x doesn't support password authentication.

To configure the Sentinel:

1. SSH in to the server that will host Consul/Sentinel.

2. Download and install the Omnibus GitLab package of your choice. Be sure to both follow only installation steps 1 and 2 on the page, and to select the correct Omnibus GitLab package, with the same version and type (Community or Enterprise editions) as your current install.

3. Edit /etc/gitlab/gitlab.rb and add the contents:

   roles(['redis_sentinel_role', 'consul_role'])
   
   # Must be the same in every sentinel node
   redis['master_name'] = 'gitlab-redis'
   
   # The same password for Redis authentication you set up for the primary node.
   redis['master_password'] = 'redis-password-goes-here'
   
   # The IP of the primary Redis node.
   redis['master_ip'] = '10.6.0.61'
   
   # Define a port so Redis can listen for TCP requests which will allow other
   # machines to connect to it.
   redis['port'] = 6379
   
   # Port of primary Redis server, uncomment to change to non default. Defaults
   # to `6379`.
   #redis['master_port'] = 6379
   
   ## Configure Sentinel
   sentinel['bind'] = '10.6.0.11'
   
   # Port that Sentinel listens on, uncomment to change to non default. Defaults
   # to `26379`.
   # sentinel['port'] = 26379
   
   ## Quorum must reflect the amount of voting sentinels it take to start a failover.
   ## Value must NOT be greater then the amount of sentinels.
   ##
   ## The quorum can be used to tune Sentinel in two ways:
   ## 1. If a the quorum is set to a value smaller than the majority of Sentinels
   ##    we deploy, we are basically making Sentinel more sensible to primary failures,
   ##    triggering a failover as soon as even just a minority of Sentinels is no longer
   ##    able to talk with the primary.
   ## 1. If a quorum is set to a value greater than the majority of Sentinels, we are
   ##    making Sentinel able to failover only when there are a very large number (larger
   ##    than majority) of well connected Sentinels which agree about the primary being down.s
   sentinel['quorum'] = 2
   
   ## Consider unresponsive server down after x amount of ms.
   # sentinel['down_after_milliseconds'] = 10000
   
   ## Specifies the failover timeout in milliseconds. It is used in many ways:
   ##
   ## - The time needed to re-start a failover after a previous failover was
   ##   already tried against the same primary by a given Sentinel, is two
   ##   times the failover timeout.
   ##
   ## - The time needed for a replica replicating to a wrong primary according
   ##   to a Sentinel current configuration, to be forced to replicate
   ##   with the right primary, is exactly the failover timeout (counting since
   ##   the moment a Sentinel detected the misconfiguration).
   ##
   ## - The time needed to cancel a failover that is already in progress but
   ##   did not produced any configuration change (REPLICAOF NO ONE yet not
   ##   acknowledged by the promoted replica).
   ##
   ## - The maximum time a failover in progress waits for all the replica to be
   ##   reconfigured as replicas of the new primary. However even after this time
   ##   the replicas will be reconfigured by the Sentinels anyway, but not with
   ##   the exact parallel-syncs progression as specified.
   # sentinel['failover_timeout'] = 60000
   
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      server: true,
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   
   # Set the network addresses that the exporters will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   redis_exporter['listen_address'] = '0.0.0.0:9121'
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. Reconfigure Omnibus GitLab for the changes to take effect.

6. Go through the steps again for all the other Consul/Sentinel nodes, and make sure you set up the correct IPs.

A Consul leader is elected when the provisioning of the third Consul server is complete. Viewing the Consul logs sudo gitlab-ctl tail consul displays ...[INFO] consul: New leader elected: ....

You can list the current Consul members (server, client):

sudo /opt/gitlab/embedded/bin/consul members

You can verify the GitLab services are running:

sudo gitlab-ctl status

The output should be similar to the following:

run: consul: (pid 30074) 76834s; run: log: (pid 29740) 76844s
run: logrotate: (pid 30925) 3041s; run: log: (pid 29649) 76861s
run: node-exporter: (pid 30093) 76833s; run: log: (pid 29663) 76855s
run: sentinel: (pid 30098) 76832s; run: log: (pid 29704) 76850s

Back to setup components

Configure PostgreSQL

In this section, you'll be guided through configuring an external PostgreSQL database to be used with GitLab.

Provide your own PostgreSQL instance

If you're hosting GitLab on a cloud provider, you can optionally use a managed service for PostgreSQL. For example, AWS offers a managed Relational Database Service (RDS) that runs PostgreSQL.

If you use a cloud-managed service, or provide your own PostgreSQL:

1. Set up PostgreSQL according to the database requirements document.
2. Set up a gitlab username with a password of your choice. The gitlab user needs privileges to create the gitlabhq_production database.
3. Configure the GitLab application servers with the appropriate details. This step is covered in Configuring the GitLab Rails application.

See Configure GitLab using an external PostgreSQL service for further configuration steps.

Standalone PostgreSQL using Omnibus GitLab

The following IPs will be used as an example:

• 10.6.0.31: PostgreSQL primary
• 10.6.0.32: PostgreSQL secondary 1
• 10.6.0.33: PostgreSQL secondary 2

First, make sure to install the Linux GitLab package on each node. Following the steps, install the necessary dependencies from step 1, and add the GitLab package repository from step 2. When installing GitLab in the second step, do not supply the EXTERNAL_URL value.

PostgreSQL nodes

1. SSH in to one of the PostgreSQL nodes.

2. Generate a password hash for the PostgreSQL username/password pair. This assumes you will use the default username of gitlab (recommended). The command will request a password and confirmation. Use the value that is output by this command in the next step as the value of <postgresql_password_hash>:

   sudo gitlab-ctl pg-password-md5 gitlab
   

3. Generate a password hash for the PgBouncer username/password pair. This assumes you will use the default username of pgbouncer (recommended). The command will request a password and confirmation. Use the value that is output by this command in the next step as the value of <pgbouncer_password_hash>:

   sudo gitlab-ctl pg-password-md5 pgbouncer
   

4. Generate a password hash for the PostgreSQL replication username/password pair. This assumes you will use the default username of gitlab_replicator (recommended). The command will request a password and a confirmation. Use the value that is output by this command in the next step as the value of <postgresql_replication_password_hash>:

   sudo gitlab-ctl pg-password-md5 gitlab_replicator
   

5. Generate a password hash for the Consul database username/password pair. This assumes you will use the default username of gitlab-consul (recommended). The command will request a password and confirmation. Use the value that is output by this command in the next step as the value of <consul_password_hash>:

   sudo gitlab-ctl pg-password-md5 gitlab-consul
   

6. On every database node, edit /etc/gitlab/gitlab.rb replacing values noted in the # START user configuration section:

   # Disable all components except Patroni and Consul
   roles(['patroni_role'])
   
   # PostgreSQL configuration
   postgresql['listen_address'] = '0.0.0.0'
   
   # Sets `max_replication_slots` to double the number of database nodes.
   # Patroni uses one extra slot per node when initiating the replication.
   patroni['postgresql']['max_replication_slots'] = 6
   
   # Set `max_wal_senders` to one more than the number of replication slots in the cluster.
   # This is used to prevent replication from using up all of the
   # available database connections.
   patroni['postgresql']['max_wal_senders'] = 7
   
   # Incoming recommended value for max connections is 500. See https://gitlab.com/gitlab-org/omnibus-gitlab/-/issues/5691.
   patroni['postgresql']['max_connections'] = 500
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   
   # Configure the Consul agent
   consul['services'] = %w(postgresql)
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   # START user configuration
   # Please set the real values as explained in Required Information section
   #
   # Replace PGBOUNCER_PASSWORD_HASH with a generated md5 value
   postgresql['pgbouncer_user_password'] = '<pgbouncer_password_hash>'
   # Replace POSTGRESQL_REPLICATION_PASSWORD_HASH with a generated md5 value
   postgresql['sql_replication_password'] = '<postgresql_replication_password_hash>'
   # Replace POSTGRESQL_PASSWORD_HASH with a generated md5 value
   postgresql['sql_user_password'] = '<postgresql_password_hash>'
   
   # Set up basic authentication for the Patroni API (use the same username/password in all nodes).
   patroni['username'] = '<patroni_api_username>'
   patroni['password'] = '<patroni_api_password>'
   
   # Replace 10.6.0.0/24 with Network Addresses for your other patroni nodes
   patroni['allowlist'] = %w(10.6.0.0/24 127.0.0.1/32)
   
   # Replace 10.6.0.0/24 with Network Address
   postgresql['trust_auth_cidr_addresses'] = %w(10.6.0.0/24 127.0.0.1/32)
   
   # Set the network addresses that the exporters will listen on for monitoring
   node_exporter['listen_address'] = '0.0.0.0:9100'
   postgres_exporter['listen_address'] = '0.0.0.0:9187'
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   #
   # END user configuration
   

PostgreSQL, with Patroni managing its failover, will default to use pg_rewind by default to handle conflicts. Like most failover handling methods, this has a small chance of leading to data loss. Learn more about the various Patroni replication methods.

1. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

2. Reconfigure GitLab for the changes to take effect.

Advanced configuration options are supported and can be added if needed.

Back to setup components

PostgreSQL post-configuration

SSH in to any of the Patroni nodes on the primary site:

1. Check the status of the leader and cluster:

   gitlab-ctl patroni members
   

   The output should be similar to the following:

   | Cluster       | Member                            |  Host     | Role   | State   | TL  | Lag in MB | Pending restart |
   |---------------|-----------------------------------|-----------|--------|---------|-----|-----------|-----------------|
   | postgresql-ha | <PostgreSQL primary hostname>     | 10.6.0.31 | Leader | running | 175 |           | *               |
   | postgresql-ha | <PostgreSQL secondary 1 hostname> | 10.6.0.32 |        | running | 175 | 0         | *               |
   | postgresql-ha | <PostgreSQL secondary 2 hostname> | 10.6.0.33 |        | running | 175 | 0         | *               |
   

If the 'State' column for any node doesn't say "running", check the Troubleshooting section before proceeding.

Back to setup components

Configure PgBouncer

Now that the PostgreSQL servers are all set up, let's configure PgBouncer. The following IPs will be used as an example:

• 10.6.0.21: PgBouncer 1
• 10.6.0.22: PgBouncer 2
• 10.6.0.23: PgBouncer 3

1. On each PgBouncer node, edit /etc/gitlab/gitlab.rb, and replace <consul_password_hash> and <pgbouncer_password_hash> with the password hashes you set up previously:

   # Disable all components except Pgbouncer and Consul agent
   roles(['pgbouncer_role'])
   
   # Configure PgBouncer
   pgbouncer['admin_users'] = %w(pgbouncer gitlab-consul)
   pgbouncer['users'] = {
      'gitlab-consul': {
         password: '<consul_password_hash>'
      },
      'pgbouncer': {
         password: '<pgbouncer_password_hash>'
      }
   }
   # Incoming recommended value for max db connections is 150. See https://gitlab.com/gitlab-org/omnibus-gitlab/-/issues/5691.
   pgbouncer['max_db_connections'] = 150
   
   # Configure Consul agent
   consul['watchers'] = %w(postgresql)
   consul['configuration'] = {
   retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13)
   }
   
   # Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] = true
   
   # Set the network addresses that the exporters will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   pgbouncer_exporter['listen_address'] = '0.0.0.0:9188'
   

2. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

3. Reconfigure Omnibus GitLab for the changes to take effect.

4. Create a .pgpass file so Consul is able to reload PgBouncer. Enter the PgBouncer password twice when asked:

   gitlab-ctl write-pgpass --host 127.0.0.1 --database pgbouncer --user pgbouncer --hostuser gitlab-consul
   

5. Ensure each node is talking to the current master:

   gitlab-ctl pgb-console # You will be prompted for PGBOUNCER_PASSWORD
   

   If there is an error psql: ERROR: Auth failed after typing in the password, ensure you previously generated the MD5 password hashes with the correct format. The correct format is to concatenate the password and the username: PASSWORDUSERNAME. For example, Sup3rS3cr3tpgbouncer would be the text needed to generate an MD5 password hash for the pgbouncer user.

6. Once the console prompt is available, run the following queries:

   show databases ; show clients ;
   

   The output should be similar to the following:

           name         |  host       | port |      database       | force_user | pool_size | reserve_pool | pool_mode | max_connections | current_connections
   ---------------------+-------------+------+---------------------+------------+-----------+--------------+-----------+-----------------+---------------------
    gitlabhq_production | MASTER_HOST | 5432 | gitlabhq_production |            |        20 |            0 |           |               0 |                   0
    pgbouncer           |             | 6432 | pgbouncer           | pgbouncer  |         2 |            0 | statement |               0 |                   0
   (2 rows)
   
    type |   user    |      database       |  state  |   addr         | port  | local_addr | local_port |    connect_time     |    request_time     |    ptr    | link | remote_pid | tls
   ------+-----------+---------------------+---------+----------------+-------+------------+------------+---------------------+---------------------+-----------+------+------------+-----
    C    | pgbouncer | pgbouncer           | active  | 127.0.0.1      | 56846 | 127.0.0.1  |       6432 | 2017-08-21 18:09:59 | 2017-08-21 18:10:48 | 0x22b3880 |      |          0 |
   (2 rows)
   

7. Verify the GitLab services are running:

   sudo gitlab-ctl status
   

   The output should be similar to the following:

   run: consul: (pid 31530) 77150s; run: log: (pid 31106) 77182s
   run: logrotate: (pid 32613) 3357s; run: log: (pid 30107) 77500s
   run: node-exporter: (pid 31550) 77149s; run: log: (pid 30138) 77493s
   run: pgbouncer: (pid 32033) 75593s; run: log: (pid 31117) 77175s
   run: pgbouncer-exporter: (pid 31558) 77148s; run: log: (pid 31498) 77156s
   

Back to setup components

Configure Gitaly Cluster

Gitaly Cluster is a GitLab provided and recommended fault tolerant solution for storing Git repositories. In this configuration, every Git repository is stored on every Gitaly node in the cluster, with one being designated the primary, and failover occurs automatically if the primary node goes down.

NOTE: Gitaly Cluster provides the benefits of fault tolerance, but comes with additional complexity of setup and management. Please review the existing technical limitations and considerations prior to deploying Gitaly Cluster. For implementations with Gitaly Sharded, the same Gitaly specs should be used. Follow the separate Gitaly documentation instead of this section.

The recommended cluster setup includes the following components:

• 3 Gitaly nodes: Replicated storage of Git repositories.
• 3 Praefect nodes: Router and transaction manager for Gitaly Cluster.
• 1 Praefect PostgreSQL node: Database server for Praefect. A third-party solution is required for Praefect database connections to be made highly available.
• 1 load balancer: A load balancer is required for Praefect. The internal load balancer will be used.

This section will detail how to configure the recommended standard setup in order. For more advanced setups refer to the standalone Gitaly Cluster documentation.

Configure Praefect PostgreSQL

Praefect, the routing and transaction manager for Gitaly Cluster, requires its own database server to store data on Gitaly Cluster status.

If you want to have a highly available setup, Praefect requires a third-party PostgreSQL database. A built-in solution is being worked on.

Praefect non-HA PostgreSQL standalone using Omnibus GitLab

The following IPs will be used as an example:

• 10.6.0.141: Praefect PostgreSQL

First, make sure to install the Linux GitLab package in the Praefect PostgreSQL node. Following the steps, install the necessary dependencies from step 1, and add the GitLab package repository from step 2. When installing GitLab in the second step, do not supply the EXTERNAL_URL value.

1. SSH in to the Praefect PostgreSQL node.

2. Create a strong password to be used for the Praefect PostgreSQL user. Take note of this password as <praefect_postgresql_password>.

3. Generate the password hash for the Praefect PostgreSQL username/password pair. This assumes you will use the default username of praefect (recommended). The command will request the password <praefect_postgresql_password> and confirmation. Use the value that is output by this command in the next step as the value of <praefect_postgresql_password_hash>:

   sudo gitlab-ctl pg-password-md5 praefect
   

4. Edit /etc/gitlab/gitlab.rb replacing values noted in the # START user configuration section:

   # Disable all components except PostgreSQL and Consul
   roles(['postgres_role', 'consul_role'])
   
   # PostgreSQL configuration
   postgresql['listen_address'] = '0.0.0.0'
   postgresql['max_connections'] = 200
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   
   # Configure the Consul agent
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   # START user configuration
   # Please set the real values as explained in Required Information section
   #
   # Replace PRAEFECT_POSTGRESQL_PASSWORD_HASH with a generated md5 value
   postgresql['sql_user_password'] = "<praefect_postgresql_password_hash>"
   
   # Replace XXX.XXX.XXX.XXX/YY with Network Address
   postgresql['trust_auth_cidr_addresses'] = %w(10.6.0.0/24 127.0.0.1/32)
   
   # Set the network addresses that the exporters will listen on for monitoring
   node_exporter['listen_address'] = '0.0.0.0:9100'
   postgres_exporter['listen_address'] = '0.0.0.0:9187'
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   #
   # END user configuration
   

5. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

6. Reconfigure GitLab for the changes to take effect.

7. Follow the post configuration.

Back to setup components

Praefect HA PostgreSQL third-party solution

As noted, a third-party PostgreSQL solution for Praefect's database is recommended if aiming for full High Availability.

There are many third-party solutions for PostgreSQL HA. The solution selected must have the following to work with Praefect:

• A static IP for all connections that doesn't change on failover.
• LISTEN SQL functionality must be supported.

Examples of the above could include Google's Cloud SQL or Amazon RDS.

Once the database is set up, follow the post configuration.

Praefect PostgreSQL post-configuration

After the Praefect PostgreSQL server has been set up, you'll then need to configure the user and database for Praefect to use.

We recommend the user be named praefect and the database praefect_production, and these can be configured as standard in PostgreSQL. The password for the user is the same as the one you configured earlier as <praefect_postgresql_password>.

This is how this would work with a Omnibus GitLab PostgreSQL setup:

1. SSH in to the Praefect PostgreSQL node.

2. Connect to the PostgreSQL server with administrative access. The gitlab-psql user should be used here for this as it's added by default in Omnibus. The database template1 is used because it is created by default on all PostgreSQL servers.

   /opt/gitlab/embedded/bin/psql -U gitlab-psql -d template1 -h POSTGRESQL_SERVER_ADDRESS
   

3. Create the new user praefect, replacing <praefect_postgresql_password>:

   CREATE ROLE praefect WITH LOGIN CREATEDB PASSWORD <praefect_postgresql_password>;
   

4. Reconnect to the PostgreSQL server, this time as the praefect user:

   /opt/gitlab/embedded/bin/psql -U praefect -d template1 -h POSTGRESQL_SERVER_ADDRESS
   

5. Create a new database praefect_production:

   CREATE DATABASE praefect_production WITH ENCODING=UTF8;
   

Back to setup components

Configure Praefect

Praefect is the router and transaction manager for Gitaly Cluster and all connections to Gitaly go through it. This section details how to configure it.

Praefect requires several secret tokens to secure communications across the Cluster:

• <praefect_external_token>: Used for repositories hosted on your Gitaly cluster and can only be accessed by Gitaly clients that carry this token.
• <praefect_internal_token>: Used for replication traffic inside your Gitaly cluster. This is distinct from praefect_external_token because Gitaly clients must not be able to access internal nodes of the Praefect cluster directly; that could lead to data loss.
• <praefect_postgresql_password>: The Praefect PostgreSQL password defined in the previous section is also required as part of this setup.

Gitaly Cluster nodes are configured in Praefect via a virtual storage. Each storage contains the details of each Gitaly node that makes up the cluster. Each storage is also given a name and this name is used in several areas of the configuration. In this guide, the name of the storage will be default. Also, this guide is geared towards new installs, if upgrading an existing environment to use Gitaly Cluster, you may need to use a different name. Refer to the Praefect documentation for more information.

The following IPs will be used as an example:

• 10.6.0.131: Praefect 1
• 10.6.0.132: Praefect 2
• 10.6.0.133: Praefect 3

To configure the Praefect nodes, on each one:

1. SSH in to the Praefect server.

2. Download and install the Omnibus GitLab package of your choice. Be sure to follow only installation steps 1 and 2 on the page.

3. Edit the /etc/gitlab/gitlab.rb file to configure Praefect:

   # Avoid running unnecessary services on the Praefect server
   gitaly['enable'] = false
   postgresql['enable'] = false
   redis['enable'] = false
   puma['enable'] = false
   sidekiq['enable'] = false
   gitlab_workhorse['enable'] = false
   prometheus['enable'] = false
   alertmanager['enable'] = false
   grafana['enable'] = false
   gitlab_exporter['enable'] = false
   nginx['enable'] = false
   
   # Praefect Configuration
   praefect['enable'] = true
   praefect['listen_addr'] = '0.0.0.0:2305'
   
   # Prevent database migrations from running on upgrade automatically
   praefect['auto_migrate'] = false
   gitlab_rails['auto_migrate'] = false
   
   # Configure the Consul agent
   consul['enable'] = true
   ## Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   
   # START user configuration
   # Please set the real values as explained in Required Information section
   #
   
   # Praefect External Token
   # This is needed by clients outside the cluster (like GitLab Shell) to communicate with the Praefect cluster
   praefect['auth_token'] = '<praefect_external_token>'
   
   # Praefect Database Settings
   praefect['database_host'] = '10.6.0.141'
   praefect['database_port'] = 5432
   # `no_proxy` settings must always be a direct connection for caching
   praefect['database_host_no_proxy'] = '10.6.0.141'
   praefect['database_port_no_proxy'] = 5432
   praefect['database_dbname'] = 'praefect_production'
   praefect['database_user'] = 'praefect'
   praefect['database_password'] = '<praefect_postgresql_password>'
   
   # Praefect Virtual Storage config
   # Name of storage hash must match storage name in git_data_dirs on GitLab
   # server ('praefect') and in git_data_dirs on Gitaly nodes ('gitaly-1')
   praefect['virtual_storages'] = {
     'default' => {
       'nodes' => {
         'gitaly-1' => {
           'address' => 'tcp://10.6.0.91:8075',
           'token'   => '<praefect_internal_token>'
         },
         'gitaly-2' => {
           'address' => 'tcp://10.6.0.92:8075',
           'token'   => '<praefect_internal_token>'
         },
         'gitaly-3' => {
           'address' => 'tcp://10.6.0.93:8075',
           'token'   => '<praefect_internal_token>'
         },
       }
     }
   }
   
   # Set the network addresses that the exporters will listen on for monitoring
   node_exporter['listen_address'] = '0.0.0.0:9100'
   praefect['prometheus_listen_addr'] = '0.0.0.0:9652'
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   #
   # END user configuration
   

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. Praefect requires to run some database migrations, much like the main GitLab application. For this you should select one Praefect node only to run the migrations, AKA the Deploy Node. This node must be configured first before the others as follows:

   1. In the /etc/gitlab/gitlab.rb file, change the praefect['auto_migrate'] setting value from false to true

   2. To ensure database migrations are only run during reconfigure and not automatically on upgrade, run:

   sudo touch /etc/gitlab/skip-auto-reconfigure
   

   1. Reconfigure GitLab for the changes to take effect and to run the Praefect database migrations.

6. On all other Praefect nodes, Reconfigure GitLab for the changes to take effect.

Configure Gitaly

The Gitaly server nodes that make up the cluster have requirements that are dependent on data, specifically the number of projects and those projects' sizes. It's recommended that a Gitaly Cluster stores no more than 5 TB of data on each node. Depending on your repository storage requirements, you may require additional Gitaly Clusters.

Due to Gitaly having notable input and output requirements, we strongly recommend that all Gitaly nodes use solid-state drives (SSDs). These SSDs should have a throughput of at least 8,000 input/output operations per second (IOPS) for read operations and 2,000 IOPS for write operations. These IOPS values are initial recommendations, and may be adjusted to greater or lesser values depending on the scale of your environment's workload. If you're running the environment on a Cloud provider, refer to their documentation about how to configure IOPS correctly.

Gitaly servers must not be exposed to the public internet, as Gitaly's network traffic is unencrypted by default. The use of a firewall is highly recommended to restrict access to the Gitaly server. Another option is to use TLS.

For configuring Gitaly you should note the following:

• git_data_dirs should be configured to reflect the storage path for the specific Gitaly node
• auth_token should be the same as praefect_internal_token

The following IPs will be used as an example:

• 10.6.0.91: Gitaly 1
• 10.6.0.92: Gitaly 2
• 10.6.0.93: Gitaly 3

On each node:

1. Download and install the Omnibus GitLab package of your choice. Be sure to follow only installation steps 1 and 2 on the page, and do not provide the EXTERNAL_URL value.

2. Edit the Gitaly server node's /etc/gitlab/gitlab.rb file to configure storage paths, enable the network listener, and to configure the token:

   # Avoid running unnecessary services on the Gitaly server
   postgresql['enable'] = false
   redis['enable'] = false
   puma['enable'] = false
   sidekiq['enable'] = false
   gitlab_workhorse['enable'] = false
   prometheus['enable'] = false
   alertmanager['enable'] = false
   grafana['enable'] = false
   gitlab_exporter['enable'] = false
   nginx['enable'] = false
   
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   
   # Configure the gitlab-shell API callback URL. Without this, `git push` will
   # fail. This can be your 'front door' GitLab URL or an internal load
   # balancer.
   gitlab_rails['internal_api_url'] = 'https://gitlab.example.com'
   
   # Gitaly
   gitaly['enable'] = true
   
   # Make Gitaly accept connections on all network interfaces. You must use
   # firewalls to restrict access to this address/port.
   # Comment out following line if you only want to support TLS connections
   gitaly['listen_addr'] = "0.0.0.0:8075"
   
   # Gitaly Auth Token
   # Should be the same as praefect_internal_token
   gitaly['auth_token'] = '<praefect_internal_token>'
   

3. Append the following to /etc/gitlab/gitlab.rb for each respective server:

   • On Gitaly node 1:

     git_data_dirs({
       "gitaly-1" => {
         "path" => "/var/opt/gitlab/git-data"
        }
     })
     

   • On Gitaly node 2:

     git_data_dirs({
       "gitaly-2" => {
         "path" => "/var/opt/gitlab/git-data"
        }
     })
     

   • On Gitaly node 3:

     git_data_dirs({
       "gitaly-3" => {
         "path" => "/var/opt/gitlab/git-data"
        }
     })
     

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. Save the file, and then reconfigure GitLab.

Gitaly Cluster TLS support

Praefect supports TLS encryption. To communicate with a Praefect instance that listens for secure connections, you must:

• Use a tls:// URL scheme in the gitaly_address of the corresponding storage entry in the GitLab configuration.
• Bring your own certificates because this isn't provided automatically. The certificate corresponding to each Praefect server must be installed on that Praefect server.

Additionally the certificate, or its certificate authority, must be installed on all Gitaly servers and on all Praefect clients that communicate with it following the procedure described in GitLab custom certificate configuration (and repeated below).

Note the following:

• The certificate must specify the address you use to access the Praefect server. If addressing the Praefect server by:

  • Hostname, you can either use the Common Name field for this, or add it as a Subject Alternative Name.
  • IP address, you must add it as a Subject Alternative Name to the certificate.

• You can configure Praefect servers with both an unencrypted listening address listen_addr and an encrypted listening address tls_listen_addr at the same time. This allows you to do a gradual transition from unencrypted to encrypted traffic, if necessary.

• The Internal Load Balancer will also access to the certificates and need to be configured to allow for TLS passthrough. Refer to the load balancers documentation on how to configure this.

To configure Praefect with TLS:

1. Create certificates for Praefect servers.

2. On the Praefect servers, create the /etc/gitlab/ssl directory and copy your key and certificate there:

   sudo mkdir -p /etc/gitlab/ssl
   sudo chmod 755 /etc/gitlab/ssl
   sudo cp key.pem cert.pem /etc/gitlab/ssl/
   sudo chmod 644 key.pem cert.pem
   

3. Edit /etc/gitlab/gitlab.rb and add:

   praefect['tls_listen_addr'] = "0.0.0.0:3305"
   praefect['certificate_path'] = "/etc/gitlab/ssl/cert.pem"
   praefect['key_path'] = "/etc/gitlab/ssl/key.pem"
   

4. Save the file and reconfigure.

5. On the Praefect clients (including each Gitaly server), copy the certificates, or their certificate authority, into /etc/gitlab/trusted-certs:

   sudo cp cert.pem /etc/gitlab/trusted-certs/
   

6. On the Praefect clients (except Gitaly servers), edit git_data_dirs in /etc/gitlab/gitlab.rb as follows:

   git_data_dirs({
     "default" => {
       "gitaly_address" => 'tls://LOAD_BALANCER_SERVER_ADDRESS:2305',
       "gitaly_token" => 'PRAEFECT_EXTERNAL_TOKEN'
     }
   })
   

7. Save the file and reconfigure GitLab.

Back to setup components

Configure Sidekiq

Sidekiq requires connection to the Redis, PostgreSQL and Gitaly instances. Object storage is also required to be configured.

• 10.6.0.71: Sidekiq 1
• 10.6.0.72: Sidekiq 2
• 10.6.0.73: Sidekiq 3
• 10.6.0.74: Sidekiq 4

To configure the Sidekiq nodes, one each one:

1. SSH in to the Sidekiq server.

2. Download and install the Omnibus GitLab package of your choice. Be sure to follow only installation steps 1 and 2 on the page.

3. Create or edit /etc/gitlab/gitlab.rb and use the following configuration:

   # Avoid running unnecessary services on the Sidekiq server
   gitaly['enable'] = false
   postgresql['enable'] = false
   redis['enable'] = false
   puma['enable'] = false
   gitlab_workhorse['enable'] = false
   prometheus['enable'] = false
   alertmanager['enable'] = false
   grafana['enable'] = false
   gitlab_exporter['enable'] = false
   nginx['enable'] = false
   
   # External URL
   ## This should match the URL of the external load balancer
   external_url 'https://gitlab.example.com'
   
   # Redis
   ## Must be the same in every sentinel node
   redis['master_name'] = 'gitlab-redis'
   
   ## The same password for Redis authentication you set up for the master node.
   redis['master_password'] = '<redis_primary_password>'
   
   ## A list of sentinels with `host` and `port`
   gitlab_rails['redis_sentinels'] = [
      {'host' => '10.6.0.11', 'port' => 26379},
      {'host' => '10.6.0.12', 'port' => 26379},
      {'host' => '10.6.0.13', 'port' => 26379},
   ]
   
   # Gitaly Cluster
   ## git_data_dirs get configured for the Praefect virtual storage
   ## Address is Internal Load Balancer for Praefect
   ## Token is praefect_external_token
   git_data_dirs({
     "default" => {
       "gitaly_address" => "tcp://10.6.0.40:2305", # internal load balancer IP
       "gitaly_token" => '<praefect_external_token>'
     }
   })
   
   # PostgreSQL
   gitlab_rails['db_host'] = '10.6.0.40' # internal load balancer IP
   gitlab_rails['db_port'] = 6432
   gitlab_rails['db_password'] = '<postgresql_user_password>'
   gitlab_rails['db_adapter'] = 'postgresql'
   gitlab_rails['db_encoding'] = 'unicode'
   ## Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   
   # Sidekiq
   sidekiq['enable'] = true
   sidekiq['listen_address'] = "0.0.0.0"
   
   ## Set number of Sidekiq queue processes to the same number as available CPUs
   sidekiq['queue_groups'] = ['*'] * 4
   
   ## Set number of Sidekiq threads per queue process to the recommend number of 10
   sidekiq['max_concurrency'] = 10
   
   # Monitoring
   consul['enable'] = true
   consul['monitoring_service_discovery'] =  true
   
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13)
   }
   
   ## Set the network addresses that the exporters will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   
   ## Add the monitoring node's IP address to the monitoring whitelist
   gitlab_rails['monitoring_whitelist'] = ['10.6.0.81/32', '127.0.0.0/8']
   gitlab_rails['prometheus_address'] = '10.6.0.81:9090'
   
   # Object Storage
   ## This is an example for configuring Object Storage on GCP
   ## Replace this config with your chosen Object Storage provider as desired
   gitlab_rails['object_store']['connection'] = {
     'provider' => 'Google',
     'google_project' => '<gcp-project-name>',
     'google_json_key_location' => '<path-to-gcp-service-account-key>'
   }
   gitlab_rails['object_store']['objects']['artifacts']['bucket'] = "<gcp-artifacts-bucket-name>"
   gitlab_rails['object_store']['objects']['external_diffs']['bucket'] = "<gcp-external-diffs-bucket-name>"
   gitlab_rails['object_store']['objects']['lfs']['bucket'] = "<gcp-lfs-bucket-name>"
   gitlab_rails['object_store']['objects']['uploads']['bucket'] = "<gcp-uploads-bucket-name>"
   gitlab_rails['object_store']['objects']['packages']['bucket'] = "<gcp-packages-bucket-name>"
   gitlab_rails['object_store']['objects']['dependency_proxy']['bucket'] = "<gcp-dependency-proxy-bucket-name>"
   gitlab_rails['object_store']['objects']['terraform_state']['bucket'] = "<gcp-terraform-state-bucket-name>"
   
   gitlab_rails['backup_upload_connection'] = {
     'provider' => 'Google',
     'google_project' => '<gcp-project-name>',
     'google_json_key_location' => '<path-to-gcp-service-account-key>'
   }
   gitlab_rails['backup_upload_remote_directory'] = "<gcp-backups-state-bucket-name>"
   

4. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

5. To ensure database migrations are only run during reconfigure and not automatically on upgrade, run:

   sudo touch /etc/gitlab/skip-auto-reconfigure
   

   Only a single designated node should handle migrations as detailed in the GitLab Rails post-configuration section.

6. Reconfigure GitLab for the changes to take effect.

7. Verify the GitLab services are running:

   sudo gitlab-ctl status
   

   The output should be similar to the following:

   run: consul: (pid 30114) 77353s; run: log: (pid 29756) 77367s
   run: logrotate: (pid 9898) 3561s; run: log: (pid 29653) 77380s
   run: node-exporter: (pid 30134) 77353s; run: log: (pid 29706) 77372s
   run: sidekiq: (pid 30142) 77351s; run: log: (pid 29638) 77386s
   

NOTE: If you find that the environment's Sidekiq job processing is slow with long queues, more nodes can be added as required. You can also tune your Sidekiq nodes to run multiple Sidekiq processes.

Back to setup components

Configure GitLab Rails

This section describes how to configure the GitLab application (Rails) component. Object storage is also required to be configured.

On each node perform the following:

1. If you're using NFS:

   1. If necessary, install the NFS client utility packages using the following commands:

      # Ubuntu/Debian
      apt-get install nfs-common
      
      # CentOS/Red Hat
      yum install nfs-utils nfs-utils-lib
      

   2. Specify the necessary NFS mounts in /etc/fstab. The exact contents of /etc/fstab will depend on how you chose to configure your NFS server. See the NFS documentation for examples and the various options.

   3. Create the shared directories. These may be different depending on your NFS mount locations.

      mkdir -p /var/opt/gitlab/.ssh /var/opt/gitlab/gitlab-rails/uploads /var/opt/gitlab/gitlab-rails/shared /var/opt/gitlab/gitlab-ci/builds /var/opt/gitlab/git-data
      

2. Download and install the Omnibus GitLab package of your choice. Be sure to follow only installation steps 1 and 2 on the page.

3. Create or edit /etc/gitlab/gitlab.rb and use the following configuration. To maintain uniformity of links across nodes, the external_url on the application server should point to the external URL that users will use to access GitLab. This would be the URL of the external load balancer which will route traffic to the GitLab application server:

   external_url 'https://gitlab.example.com'
   
   # git_data_dirs get configured for the Praefect virtual storage
   # Address is Interal Load Balancer for Praefect
   # Token is praefect_external_token
   git_data_dirs({
     "default" => {
       "gitaly_address" => "tcp://10.6.0.40:2305", # internal load balancer IP
       "gitaly_token" => '<praefect_external_token>'
     }
   })
   
   ## Disable components that will not be on the GitLab application server
   roles(['application_role'])
   gitaly['enable'] = false
   nginx['enable'] = true
   sidekiq['enable'] = false
   
   ## PostgreSQL connection details
   # Disable PostgreSQL on the application node
   postgresql['enable'] = false
   gitlab_rails['db_host'] = '10.6.0.20' # internal load balancer IP
   gitlab_rails['db_port'] = 6432
   gitlab_rails['db_password'] = '<postgresql_user_password>'
   # Prevent database migrations from running on upgrade automatically
   gitlab_rails['auto_migrate'] = false
   
   ## Redis connection details
   ## Must be the same in every sentinel node
   redis['master_name'] = 'gitlab-redis'
   
   ## The same password for Redis authentication you set up for the Redis primary node.
   redis['master_password'] = '<redis_primary_password>'
   
   ## A list of sentinels with `host` and `port`
   gitlab_rails['redis_sentinels'] = [
     {'host' => '10.6.0.11', 'port' => 26379},
     {'host' => '10.6.0.12', 'port' => 26379},
     {'host' => '10.6.0.13', 'port' => 26379}
   ]
   
   ## Enable service discovery for Prometheus
   consul['enable'] = true
   consul['monitoring_service_discovery'] =  true
   
   # Set the network addresses that the exporters used for monitoring will listen on
   node_exporter['listen_address'] = '0.0.0.0:9100'
   gitlab_workhorse['prometheus_listen_addr'] = '0.0.0.0:9229'
   sidekiq['listen_address'] = "0.0.0.0"
   puma['listen'] = '0.0.0.0'
   
   ## The IPs of the Consul server nodes
   ## You can also use FQDNs and intermix them with IPs
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13),
   }
   
   # Add the monitoring node's IP address to the monitoring whitelist and allow it to
   # scrape the NGINX metrics
   gitlab_rails['monitoring_whitelist'] = ['10.6.0.81/32', '127.0.0.0/8']
   nginx['status']['options']['allow'] = ['10.6.0.81/32', '127.0.0.0/8']
   gitlab_rails['prometheus_address'] = '10.6.0.81:9090'
   
   #############################
   ###     Object storage    ###
   #############################
   
   # This is an example for configuring Object Storage on GCP
   # Replace this config with your chosen Object Storage provider as desired
   gitlab_rails['object_store']['connection'] = {
     'provider' => 'Google',
     'google_project' => '<gcp-project-name>',
     'google_json_key_location' => '<path-to-gcp-service-account-key>'
   }
   gitlab_rails['object_store']['objects']['artifacts']['bucket'] = "<gcp-artifacts-bucket-name>"
   gitlab_rails['object_store']['objects']['external_diffs']['bucket'] = "<gcp-external-diffs-bucket-name>"
   gitlab_rails['object_store']['objects']['lfs']['bucket'] = "<gcp-lfs-bucket-name>"
   gitlab_rails['object_store']['objects']['uploads']['bucket'] = "<gcp-uploads-bucket-name>"
   gitlab_rails['object_store']['objects']['packages']['bucket'] = "<gcp-packages-bucket-name>"
   gitlab_rails['object_store']['objects']['dependency_proxy']['bucket'] = "<gcp-dependency-proxy-bucket-name>"
   gitlab_rails['object_store']['objects']['terraform_state']['bucket'] = "<gcp-terraform-state-bucket-name>"
   
   gitlab_rails['backup_upload_connection'] = {
     'provider' => 'Google',
     'google_project' => '<gcp-project-name>',
     'google_json_key_location' => '<path-to-gcp-service-account-key>'
   }
   gitlab_rails['backup_upload_remote_directory'] = "<gcp-backups-state-bucket-name>"
   
   ## Uncomment and edit the following options if you have set up NFS
   ##
   ## Prevent GitLab from starting if NFS data mounts are not available
   ##
   #high_availability['mountpoint'] = '/var/opt/gitlab/git-data'
   ##
   ## Ensure UIDs and GIDs match between servers for permissions via NFS
   ##
   #user['uid'] = 9000
   #user['gid'] = 9000
   #web_server['uid'] = 9001
   #web_server['gid'] = 9001
   #registry['uid'] = 9002
   #registry['gid'] = 9002
   

4. If you're using Gitaly with TLS support, make sure the git_data_dirs entry is configured with tls instead of tcp:

   git_data_dirs({
     "default" => {
       "gitaly_address" => "tls://10.6.0.40:2305", # internal load balancer IP
       "gitaly_token" => '<praefect_external_token>'
     }
   })
   

   1. Copy the cert into /etc/gitlab/trusted-certs:

      sudo cp cert.pem /etc/gitlab/trusted-certs/
      

5. Copy the /etc/gitlab/gitlab-secrets.json file from the first Omnibus node you configured and add or replace the file of the same name on this server. If this is the first Omnibus node you are configuring then you can skip this step.

6. To ensure database migrations are only run during reconfigure and not automatically on upgrade, run:

   sudo touch /etc/gitlab/skip-auto-reconfigure
   

   Only a single designated node should handle migrations as detailed in the GitLab Rails post-configuration section.

7. Reconfigure GitLab for the changes to take effect.

8. Run sudo gitlab-rake gitlab:gitaly:check to confirm the node can connect to Gitaly.

9. Tail the logs to see the requests:

   sudo gitlab-ctl tail gitaly
   

10. Verify the GitLab services are running:

    sudo gitlab-ctl status
    

    The output should be similar to the following:

    run: consul: (pid 4890) 8647s; run: log: (pid 29962) 79128s
    run: gitlab-exporter: (pid 4902) 8647s; run: log: (pid 29913) 79134s
    run: gitlab-workhorse: (pid 4904) 8646s; run: log: (pid 29713) 79155s
    run: logrotate: (pid 12425) 1446s; run: log: (pid 29798) 79146s
    run: nginx: (pid 4925) 8646s; run: log: (pid 29726) 79152s
    run: node-exporter: (pid 4931) 8645s; run: log: (pid 29855) 79140s
    run: puma: (pid 4936) 8645s; run: log: (pid 29656) 79161s
    

When you specify https in the external_url, as in the previous example, GitLab expects that the SSL certificates are in /etc/gitlab/ssl/. If the certificates aren't present, NGINX will fail to start. For more information, see the NGINX documentation.

GitLab Rails post-configuration

1. Ensure that all migrations ran:

   gitlab-rake gitlab:db:configure
   

   If you encounter a rake aborted! error message stating that PgBouncer is failing to connect to PostgreSQL, it may be that your PgBouncer node's IP address is missing from PostgreSQL's trust_auth_cidr_addresses in gitlab.rb on your database nodes. Before proceeding, see PgBouncer error ERROR: pgbouncer cannot connect to server.

2. Configure fast lookup of authorized SSH keys in the database.

Back to setup components

Configure Prometheus

The Omnibus GitLab package can be used to configure a standalone Monitoring node running Prometheus and Grafana:

1. SSH in to the Monitoring node.

2. Download and install the Omnibus GitLab package of your choice. Be sure to follow only installation steps 1 and 2 on the page.

3. Edit /etc/gitlab/gitlab.rb and add the contents:

   roles(['monitoring_role', 'consul_role'])
   
   external_url 'http://gitlab.example.com'
   
   # Prometheus
   prometheus['listen_address'] = '0.0.0.0:9090'
   prometheus['monitor_kubernetes'] = false
   
   # Grafana
   grafana['admin_password'] = '<grafana_password>'
   grafana['disable_login_form'] = false
   
   # Enable service discovery for Prometheus
   consul['monitoring_service_discovery'] =  true
   consul['configuration'] = {
      retry_join: %w(10.6.0.11 10.6.0.12 10.6.0.13)
   }
   
   # Nginx - For Grafana access
   nginx['enable'] = true
   

4. Save the file and reconfigure GitLab.

5. In the GitLab UI, set admin/application_settings/metrics_and_profiling > Metrics - Grafana to /-/grafana to http[s]://<MONITOR NODE>/-/grafana.

6. Verify the GitLab services are running:

   sudo gitlab-ctl status
   

   The output should be similar to the following:

   run: consul: (pid 31637) 17337s; run: log: (pid 29748) 78432s
   run: grafana: (pid 31644) 17337s; run: log: (pid 29719) 78438s
   run: logrotate: (pid 31809) 2936s; run: log: (pid 29581) 78462s
   run: nginx: (pid 31665) 17335s; run: log: (pid 29556) 78468s
   run: prometheus: (pid 31672) 17335s; run: log: (pid 29633) 78456s
   

Back to setup components

Configure the object storage

GitLab supports using an object storage service for holding numerous types of data. It's recommended over NFS and in general it's better in larger setups as object storage is typically much more performant, reliable, and scalable.

GitLab has been tested on a number of object storage providers:

• Amazon S3
• Google Cloud Storage
• Digital Ocean Spaces
• Oracle Cloud Infrastructure
• OpenStack Swift
• Azure Blob storage
• On-premises hardware and appliances from various storage vendors.
• MinIO. We have a guide to deploying this within our Helm Chart documentation.

There are two ways of specifying object storage configuration in GitLab:

• Consolidated form: A single credential is shared by all supported object types.
• Storage-specific form: Every object defines its own object storage connection and configuration.

Starting with GitLab 13.2, consolidated object storage configuration is available. It simplifies your GitLab configuration since the connection details are shared across object types. Refer to Consolidated object storage configuration guide for instructions on how to set it up.

GitLab Runner returns job logs in chunks which Omnibus GitLab caches temporarily on disk in /var/opt/gitlab/gitlab-ci/builds by default, even when using consolidated object storage. With default configuration, this directory needs to be shared via NFS on any GitLab Rails and Sidekiq nodes. In GitLab 13.6 and later, it's recommended to switch to Incremental logging, which uses Redis instead of disk space for temporary caching of job logs.

For configuring object storage in GitLab 13.1 and earlier, or for storage types not supported by consolidated configuration form, refer to the following guides based on what features you intend to use:

Object storage type	Supported by consolidated configuration?
Backups	No
Job artifacts including archived job logs	Yes
LFS objects	Yes
Uploads	Yes
Container Registry (optional feature)	No
Merge request diffs	Yes
Mattermost	No
Packages (optional feature)	Yes
Dependency Proxy (optional feature)	Yes
Pseudonymizer (optional feature)	No
Autoscale runner caching (optional for improved performance)	No
Terraform state files	Yes

Using separate buckets for each data type is the recommended approach for GitLab. This ensures there are no collisions across the various types of data GitLab stores. There are plans to enable the use of a single bucket in the future.

Back to setup components

Configure Advanced Search

You can leverage Elasticsearch and enable Advanced Search for faster, more advanced code search across your entire GitLab instance.

Elasticsearch cluster design and requirements are dependent on your specific data. For recommended best practices about how to set up your Elasticsearch cluster alongside your instance, read how to choose the optimal cluster configuration.

Back to setup components

Configure NFS (optional)

Object storage, along with Gitaly are recommended over NFS wherever possible for improved performance. If you intend to use GitLab Pages, this currently requires NFS.

See how to configure NFS.

WARNING: Engineering support for NFS for Git repositories is deprecated. Technical support is planned to be unavailable from GitLab 15.0. No further enhancements are planned for this feature.

Read:

• Gitaly and NFS Deprecation.
• About the correct mount options to use.

Cloud Native Hybrid reference architecture with Helm Charts (alternative)

As an alternative approach, you can also run select components of GitLab as Cloud Native in Kubernetes via our official Helm Charts. In this setup, we support running the equivalent of GitLab Rails and Sidekiq nodes in a Kubernetes cluster, named Webservice and Sidekiq respectively. In addition, the following other supporting services are supported: NGINX, Task Runner, Migrations, Prometheus, and Grafana.

Hybrid installations leverage the benefits of both cloud native and traditional compute deployments. With this, stateless components can benefit from cloud native workload management benefits while stateful components are deployed in compute VMs with Omnibus to benefit from increased permanence.

NOTE: This is an advanced setup. Running services in Kubernetes is well known to be complex. This setup is only recommended if you have strong working knowledge and experience in Kubernetes. The rest of this section assumes this.

Cluster topology

The following tables and diagram detail the hybrid environment using the same formats as the normal environment above.

First are the components that run in Kubernetes. The recommendation at this time is to use Google Cloud's Kubernetes Engine (GKE) and associated machine types, but the memory and CPU requirements should translate to most other providers. We hope to update this in the future with further specific cloud provider details.

Service	Nodes1	Configuration	GCP	Allocatable CPUs and Memory
Webservice	5	16 vCPU, 14.4 GB memory	n1-highcpu-16	79.5 vCPU, 62 GB memory
Sidekiq	3	4 vCPU, 15 GB memory	n1-standard-4	11.8 vCPU, 38.9 GB memory
Supporting services such as NGINX, Prometheus	2	2 vCPU, 7.5 GB memory	n1-standard-2	3.9 vCPU, 11.8 GB memory

1. Nodes configuration is shown as it is forced to ensure pod vcpu / memory ratios and avoid scaling during performance testing. In production deployments there is no need to assign pods to nodes. A minimum of three nodes in three different availability zones is strongly recommended to align with resilient cloud architecture practices.

Next are the backend components that run on static compute VMs via Omnibus (or External PaaS services where applicable):

Service	Nodes	Configuration	GCP
Redis2	3	2 vCPU, 7.5 GB memory	n1-standard-2
Consul1 + Sentinel2	3	2 vCPU, 1.8 GB memory	n1-highcpu-2
PostgreSQL1	3	4 vCPU, 15 GB memory	n1-standard-4
PgBouncer1	3	2 vCPU, 1.8 GB memory	n1-highcpu-2
Internal load balancing node3	1	2 vCPU, 1.8 GB memory	n1-highcpu-2
Gitaly5	3	8 vCPU, 30 GB memory	n1-standard-8
Praefect5	3	2 vCPU, 1.8 GB memory	n1-highcpu-2
Praefect PostgreSQL1	1+	2 vCPU, 1.8 GB memory	n1-highcpu-2
Object storage4	n/a	n/a	n/a

1. Can be optionally run on reputable third-party external PaaS PostgreSQL solutions. Google Cloud SQL and AWS RDS are known to work, however Azure Database for PostgreSQL is not recommended due to performance issues. Consul is primarily used for PostgreSQL high availability so can be ignored when using a PostgreSQL PaaS setup. However it is also used optionally by Prometheus for Omnibus auto host discovery.
2. Can be optionally run on reputable third-party external PaaS Redis solutions. Google Memorystore and AWS Elasticache are known to work.
3. Can be optionally run on reputable third-party load balancing services (LB PaaS). AWS ELB is known to work.
4. Should be run on reputable third-party object storage (storage PaaS) for cloud implementations. Google Cloud Storage and AWS S3 are known to work.
5. Gitaly Cluster provides the benefits of fault tolerance, but comes with additional complexity of setup and management. Please review the existing technical limitations and considerations prior to deploying Gitaly Cluster. If Gitaly Sharded is desired, the same specs listed above for Gitaly should be used.

NOTE: For all PaaS solutions that involve configuring instances, it is strongly recommended to implement a minimum of three nodes in three different availability zones to align with resilient cloud architecture practices.

@startuml 5k

card "Kubernetes via Helm Charts" as kubernetes {
  card "**External Load Balancer**" as elb #6a9be7

  together {
    collections "**Webservice** x5" as gitlab #32CD32
    collections "**Sidekiq** x3" as sidekiq #ff8dd1
  }

  card "**Prometheus + Grafana**" as monitor #7FFFD4
  card "**Supporting Services**" as support
}

card "**Internal Load Balancer**" as ilb #9370DB

card "**Consul + Sentinel**" as consul_sentinel {
 collections "**Consul** x3" as consul #e76a9b
 collections "**Redis Sentinel** x3" as sentinel #e6e727
}

card "Gitaly Cluster" as gitaly_cluster {
  collections "**Praefect** x3" as praefect #FF8C00
  collections "**Gitaly** x3" as gitaly #FF8C00
  card "**Praefect PostgreSQL***\n//Non fault-tolerant//" as praefect_postgres #FF8C00

  praefect -[#FF8C00]-> gitaly
  praefect -[#FF8C00]> praefect_postgres
}

card "Database" as database {
  collections "**PGBouncer** x3" as pgbouncer #4EA7FF
  card "**PostgreSQL** (Primary)" as postgres_primary #4EA7FF
  collections "**PostgreSQL** (Secondary) x2" as postgres_secondary #4EA7FF

  pgbouncer -[#4EA7FF]-> postgres_primary
  postgres_primary .[#4EA7FF]> postgres_secondary
}

card "Redis" as redis {
  collections "**Redis** x3" as redis_nodes #FF6347

  redis_nodes <.[#FF6347]- sentinel
}

cloud "**Object Storage**" as object_storage #white

elb -[#6a9be7]-> gitlab
elb -[#6a9be7]-> monitor
elb -[hidden]-> support

gitlab -[#32CD32]--> ilb
gitlab -[#32CD32]-> object_storage
gitlab -[#32CD32]---> redis
gitlab -[hidden]--> consul

sidekiq -[#ff8dd1]--> ilb
sidekiq -[#ff8dd1]-> object_storage
sidekiq -[#ff8dd1]---> redis
sidekiq -[hidden]--> consul

ilb -[#9370DB]-> gitaly_cluster
ilb -[#9370DB]-> database

consul .[#e76a9b]-> database
consul .[#e76a9b]-> gitaly_cluster
consul .[#e76a9b,norank]--> redis

monitor .[#7FFFD4]> consul
monitor .[#7FFFD4]-> database
monitor .[#7FFFD4]-> gitaly_cluster
monitor .[#7FFFD4,norank]--> redis
monitor .[#7FFFD4]> ilb
monitor .[#7FFFD4,norank]u--> elb

@enduml

Resource usage settings

The following formulas help when calculating how many pods may be deployed within resource constraints. The 5k reference architecture example values file documents how to apply the calculated configuration to the Helm Chart.

Webservice

Webservice pods typically need about 1 vCPU and 1.25 GB of memory per worker. Each Webservice pod consumes roughly 4 vCPUs and 5 GB of memory using the recommended topology because four worker processes are created by default and each pod has other small processes running.

For 5,000 users we recommend a total Puma worker count of around 40. With the provided recommendations this allows the deployment of up to 10 Webservice pods with 4 workers per pod and 2 pods per node. Expand available resources using the ratio of 1 vCPU to 1.25 GB of memory per each worker process for each additional Webservice pod.

For further information on resource usage, see the Webservice resources.

Sidekiq

Sidekiq pods should generally have 1 vCPU and 2 GB of memory.

The provided starting point allows the deployment of up to 8 Sidekiq pods. Expand available resources using the 1 vCPU to 2GB memory ratio for each additional pod.

For further information on resource usage, see the Sidekiq resources.

Back to setup components