gitlab-org--gitlab-foss/spec/models/push_event_spec.rb

309 lines
8.0 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
require 'spec_helper'
RSpec.describe PushEvent do
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
let(:payload) { PushEventPayload.new }
let(:event) do
event = described_class.new
allow(event).to receive(:push_event_payload).and_return(payload)
event
end
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
describe '.created_or_pushed' do
let(:event1) { create(:push_event) }
let(:event2) { create(:push_event) }
let(:event3) { create(:push_event) }
before do
create(:push_event_payload, event: event1, action: :pushed)
create(:push_event_payload, event: event2, action: :created)
create(:push_event_payload, event: event3, action: :removed)
end
let(:relation) { described_class.created_or_pushed }
it 'includes events for pushing to existing refs' do
expect(relation).to include(event1)
end
it 'includes events for creating new refs' do
expect(relation).to include(event2)
end
it 'does not include events for removing refs' do
expect(relation).not_to include(event3)
end
end
describe '.branch_events' do
let(:event1) { create(:push_event) }
let(:event2) { create(:push_event) }
before do
create(:push_event_payload, event: event1, ref_type: :branch)
create(:push_event_payload, event: event2, ref_type: :tag)
end
let(:relation) { described_class.branch_events }
it 'includes events for branches' do
expect(relation).to include(event1)
end
it 'does not include events for tags' do
expect(relation).not_to include(event2)
end
end
describe '.without_existing_merge_requests' do
let(:project) { create(:project, :repository) }
let(:event1) { create(:push_event, project: project) }
let(:event2) { create(:push_event, project: project) }
let(:event3) { create(:push_event, project: project) }
let(:event4) { create(:push_event, project: project) }
let(:event5) { create(:push_event, project: project) }
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
before do
create(:push_event_payload, event: event1, ref: 'foo', action: :created)
create(:push_event_payload, event: event2, ref: 'bar', action: :created)
create(:push_event_payload, event: event3, ref: 'qux', action: :created)
create(:push_event_payload, event: event4, ref: 'baz', action: :removed)
create(:push_event_payload, event: event5, ref: 'baz', ref_type: :tag)
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
project.repository.create_branch('bar')
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
create(
:merge_request,
source_project: project,
target_project: project,
source_branch: 'bar'
)
project.repository.create_branch('qux')
create(
:merge_request,
:closed,
source_project: project,
target_project: project,
source_branch: 'qux'
)
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
end
let(:relation) { described_class.without_existing_merge_requests }
it 'includes events that do not have a corresponding merge request' do
expect(relation).to include(event1)
end
it 'does not include events that have a corresponding open merge request' do
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
expect(relation).not_to include(event2)
end
it 'includes events that has corresponding closed/merged merge requests' do
expect(relation).to include(event3)
end
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
it 'does not include events for removed refs' do
expect(relation).not_to include(event4)
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
end
it 'does not include events for pushing to tags' do
expect(relation).not_to include(event5)
Rework how recent push events are retrieved Whenever you push to a branch GitLab will show a button to create a merge request (should one not exist already). The underlying code to display this data was quite inefficient. For example, it involved multiple slow queries just to figure out what the most recent push event was. This commit changes the way this data is retrieved so it's much faster. This is achieved by caching the ID of the last push event on every push, which is then retrieved when loading certain pages. Database queries are only executed if necessary and the cached data is removed automatically once a merge request has been created, or 2 hours after being stored. A trade-off of this approach is that we _only_ track the last event. Previously if you were to push to branch A and B then create a merge request for branch B we'd still show the widget for branch A. As of this commit this is no longer the case, instead we will only show the widget for the branch you pushed to most recently. Once a merge request exists the widget is no longer displayed. Alternative solutions are either too complex and/or too slow, hence the decision was made to settle for this trade-off. Performance Impact ------------------ In the best case scenario (= a user didn't push anything for more than 2 hours) we perform a single Redis GET per page. Should there be cached data we will run a single (and lightweight) SQL query to get the event data from the database. If a merge request already exists we will run an additional DEL to remove the cache key. The difference in response timings can vary a bit per project. On GitLab.com the 99th percentile of time spent in User#recent_push hovers between 100 milliseconds and 1 second, while the mean hovers around 50 milliseconds. With the changes in this MR the expected time spent in User#recent_push is expected to be reduced down to just a few milliseconds. Fixes https://gitlab.com/gitlab-org/gitlab-ce/issues/35990
2017-09-01 10:50:14 +00:00
end
end
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
describe '.sti_name' do
it 'returns the integer representation of the :pushed event action' do
expect(described_class.sti_name).to eq(Event.actions[:pushed])
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
end
end
describe '#push_action?' do
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
it 'returns true' do
expect(event).to be_push_action
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
end
end
describe '#push_with_commits?' do
it 'returns true when both the first and last commit are present' do
allow(event).to receive(:commit_from).and_return('123')
allow(event).to receive(:commit_to).and_return('456')
expect(event).to be_push_with_commits
end
it 'returns false when the first commit is missing' do
allow(event).to receive(:commit_to).and_return('456')
expect(event).not_to be_push_with_commits
end
it 'returns false when the last commit is missing' do
allow(event).to receive(:commit_from).and_return('123')
expect(event).not_to be_push_with_commits
end
end
describe '#tag?' do
it 'returns true when pushing to a tag' do
allow(payload).to receive(:tag?).and_return(true)
expect(event).to be_tag
end
it 'returns false when pushing to a branch' do
allow(payload).to receive(:tag?).and_return(false)
expect(event).not_to be_tag
end
end
describe '#branch?' do
it 'returns true when pushing to a branch' do
allow(payload).to receive(:branch?).and_return(true)
expect(event).to be_branch
end
it 'returns false when pushing to a tag' do
allow(payload).to receive(:branch?).and_return(false)
expect(event).not_to be_branch
end
end
describe '#valid_push?' do
it 'returns true if a ref exists' do
allow(payload).to receive(:ref).and_return('master')
expect(event).to be_valid_push
end
it 'returns false when no ref is present' do
expect(event).not_to be_valid_push
end
end
describe '#new_ref?' do
it 'returns true when pushing a new ref' do
allow(payload).to receive(:created?).and_return(true)
expect(event).to be_new_ref
end
it 'returns false when pushing to an existing ref' do
allow(payload).to receive(:created?).and_return(false)
expect(event).not_to be_new_ref
end
end
describe '#rm_ref?' do
it 'returns true when removing an existing ref' do
allow(payload).to receive(:removed?).and_return(true)
expect(event).to be_rm_ref
end
it 'returns false when pushing to an existing ref' do
allow(payload).to receive(:removed?).and_return(false)
expect(event).not_to be_rm_ref
end
end
describe '#commit_from' do
it 'returns the first commit SHA' do
allow(payload).to receive(:commit_from).and_return('123')
expect(event.commit_from).to eq('123')
end
end
describe '#commit_to' do
it 'returns the last commit SHA' do
allow(payload).to receive(:commit_to).and_return('123')
expect(event.commit_to).to eq('123')
end
end
describe '#ref_name' do
it 'returns the name of the ref' do
allow(payload).to receive(:ref).and_return('master')
expect(event.ref_name).to eq('master')
end
end
describe '#ref_type' do
it 'returns the type of the ref' do
allow(payload).to receive(:ref_type).and_return('branch')
expect(event.ref_type).to eq('branch')
end
end
describe '#branch_name' do
it 'returns the name of the branch' do
allow(payload).to receive(:ref).and_return('master')
expect(event.branch_name).to eq('master')
end
end
describe '#tag_name' do
it 'returns the name of the tag' do
allow(payload).to receive(:ref).and_return('1.2')
expect(event.tag_name).to eq('1.2')
end
end
describe '#commit_title' do
it 'returns the commit message' do
allow(payload).to receive(:commit_title).and_return('foo')
expect(event.commit_title).to eq('foo')
end
end
describe '#commit_id' do
it 'returns the SHA of the last commit if present' do
allow(event).to receive(:commit_to).and_return('123')
expect(event.commit_id).to eq('123')
end
it 'returns the SHA of the first commit if the last commit is not present' do
allow(event).to receive(:commit_to).and_return(nil)
allow(event).to receive(:commit_from).and_return('123')
expect(event.commit_id).to eq('123')
end
end
describe '#commits_count' do
it 'returns the number of commits' do
allow(payload).to receive(:commit_count).and_return(1)
expect(event.commits_count).to eq(1)
end
end
describe '#validate_push_action' do
it 'adds an error when the action is not PUSHED' do
event.action = :created
Migrate events into a new format This commit migrates events data in such a way that push events are stored much more efficiently. This is done by creating a shadow table called "events_for_migration", and a table called "push_event_payloads" which is used for storing push data of push events. The background migration in this commit will copy events from the "events" table into the "events_for_migration" table, push events in will also have a row created in "push_event_payloads". This approach allows us to reclaim space in the next release by simply swapping the "events" and "events_for_migration" tables, then dropping the old events (now "events_for_migration") table. The new table structure is also optimised for storage space, and does not include the unused "title" column nor the "data" column (since this data is moved to "push_event_payloads"). == Newly Created Events Newly created events are inserted into both "events" and "events_for_migration", both using the exact same primary key value. The table "push_event_payloads" in turn has a foreign key to the _shadow_ table. This removes the need for recreating and validating the foreign key after swapping the tables. Since the shadow table also has a foreign key to "projects.id" we also don't have to worry about orphaned rows. This approach however does require some additional storage as we're duplicating a portion of the events data for at least 1 release. The exact amount is hard to estimate, but for GitLab.com this is expected to be between 10 and 20 GB at most. The background migration in this commit deliberately does _not_ update the "events" table as doing so would put a lot of pressure on PostgreSQL's auto vacuuming system. == Supporting Both Old And New Events Application code has also been adjusted to support push events using both the old and new data formats. This is done by creating a PushEvent class which extends the regular Event class. Using Rails' Single Table Inheritance system we can ensure the right class is used for the right data, which in this case is based on the value of `events.action`. To support displaying old and new data at the same time the PushEvent class re-defines a few methods of the Event class, falling back to their original implementations for push events in the old format. Once all existing events have been migrated the various push event related methods can be removed from the Event model, and the calls to `super` can be removed from the methods in the PushEvent model. The UI and event atom feed have also been slightly changed to better handle this new setup, fortunately only a few changes were necessary to make this work. == API Changes The API only displays push data of events in the new format. Supporting both formats in the API is a bit more difficult compared to the UI. Since the old push data was not really well documented (apart from one example that used an incorrect "action" nmae) I decided that supporting both was not worth the effort, especially since events will be migrated in a few days _and_ new events are created in the correct format.
2017-07-10 15:43:57 +00:00
event.validate_push_action
expect(event.errors.count).to eq(1)
end
end
end