1
0
Fork 0
mirror of https://github.com/moby/moby.git synced 2022-11-09 12:21:53 -05:00
moby--moby/daemon/logger/splunk/splunk_test.go

1395 lines
36 KiB
Go
Raw Normal View History

package splunk // import "github.com/docker/docker/daemon/logger/splunk"
import (
"compress/gzip"
"context"
"fmt"
"net/http"
"os"
"runtime"
"testing"
"time"
"github.com/docker/docker/daemon/logger"
"gotest.tools/v3/assert"
"gotest.tools/v3/env"
)
// Validate options
func TestValidateLogOpt(t *testing.T) {
err := ValidateLogOpt(map[string]string{
splunkURLKey: "http://127.0.0.1",
splunkTokenKey: "2160C7EF-2CE9-4307-A180-F852B99CF417",
splunkSourceKey: "mysource",
splunkSourceTypeKey: "mysourcetype",
splunkIndexKey: "myindex",
splunkCAPathKey: "/usr/cert.pem",
splunkCANameKey: "ca_name",
splunkInsecureSkipVerifyKey: "true",
splunkFormatKey: "json",
splunkVerifyConnectionKey: "true",
splunkGzipCompressionKey: "true",
splunkGzipCompressionLevelKey: "1",
envKey: "a",
envRegexKey: "^foo",
labelsKey: "b",
tagKey: "c",
})
if err != nil {
t.Fatal(err)
}
err = ValidateLogOpt(map[string]string{
"not-supported-option": "a",
})
if err == nil {
t.Fatal("Expecting error on unsupported options")
}
}
// Driver require user to specify required options
func TestNewMissedConfig(t *testing.T) {
info := logger.Info{
Config: map[string]string{},
}
_, err := New(info)
if err == nil {
t.Fatal("Logger driver should fail when no required parameters specified")
}
}
// Driver require user to specify splunk-url
func TestNewMissedUrl(t *testing.T) {
info := logger.Info{
Config: map[string]string{
splunkTokenKey: "4642492F-D8BD-47F1-A005-0C08AE4657DF",
},
}
_, err := New(info)
if err.Error() != "splunk: splunk-url is expected" {
t.Fatal("Logger driver should fail when no required parameters specified")
}
}
// Driver require user to specify splunk-token
func TestNewMissedToken(t *testing.T) {
info := logger.Info{
Config: map[string]string{
splunkURLKey: "http://127.0.0.1:8088",
},
}
_, err := New(info)
if err.Error() != "splunk: splunk-token is expected" {
t.Fatal("Logger driver should fail when no required parameters specified")
}
}
func TestNewWithProxy(t *testing.T) {
proxy := "http://proxy.testing:8888"
reset := env.Patch(t, "HTTP_PROXY", proxy)
defer reset()
// must not be localhost
splunkURL := "http://example.com:12345"
logger, err := New(logger.Info{
Config: map[string]string{
splunkURLKey: splunkURL,
splunkTokenKey: "token",
splunkVerifyConnectionKey: "false",
},
ContainerID: "containeriid",
})
assert.NilError(t, err)
splunkLogger := logger.(*splunkLoggerInline)
proxyFunc := splunkLogger.transport.Proxy
assert.Assert(t, proxyFunc != nil)
req, err := http.NewRequest(http.MethodGet, splunkURL, nil)
assert.NilError(t, err)
proxyURL, err := proxyFunc(req)
assert.NilError(t, err)
assert.Assert(t, proxyURL != nil)
assert.Equal(t, proxy, proxyURL.String())
}
// Test default settings
func TestDefault(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
hostname, err := info.Hostname()
if err != nil {
t.Fatal(err)
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if loggerDriver.Name() != driverName {
t.Fatal("Unexpected logger driver name")
}
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerInline)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "" ||
splunkLoggerDriver.nullMessage.SourceType != "" ||
splunkLoggerDriver.nullMessage.Index != "" ||
splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize {
t.Fatal("Found not default values setup in Splunk Logging Driver.")
}
message1Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("{\"a\":\"b\"}"), Source: "stdout", Timestamp: message1Time}); err != nil {
t.Fatal(err)
}
message2Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("notajson"), Source: "stdout", Timestamp: message2Time}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 2 {
t.Fatal("Expected two messages")
}
if *hec.gzipEnabled {
t.Fatal("Gzip should not be used")
}
message1 := hec.messages[0]
if message1.Time != fmt.Sprintf("%f", float64(message1Time.UnixNano())/float64(time.Second)) ||
message1.Host != hostname ||
message1.Source != "" ||
message1.SourceType != "" ||
message1.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message1)
}
if event, err := message1.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != "{\"a\":\"b\"}" ||
event["source"] != "stdout" ||
event["tag"] != "containeriid" ||
len(event) != 3 {
t.Fatalf("Unexpected event in message %v", event)
}
}
message2 := hec.messages[1]
if message2.Time != fmt.Sprintf("%f", float64(message2Time.UnixNano())/float64(time.Second)) ||
message2.Host != hostname ||
message2.Source != "" ||
message2.SourceType != "" ||
message2.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message2)
}
if event, err := message2.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != "notajson" ||
event["source"] != "stdout" ||
event["tag"] != "containeriid" ||
len(event) != 3 {
t.Fatalf("Unexpected event in message %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify inline format with a not default settings for most of options
func TestInlineFormatWithNonDefaultOptions(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkSourceKey: "mysource",
splunkSourceTypeKey: "mysourcetype",
splunkIndexKey: "myindex",
splunkFormatKey: splunkFormatInline,
splunkGzipCompressionKey: "true",
tagKey: "{{.ImageName}}/{{.Name}}",
labelsKey: "a",
envRegexKey: "^foo",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
ContainerLabels: map[string]string{
"a": "b",
},
ContainerEnv: []string{"foo_finder=bar"},
}
hostname, err := info.Hostname()
if err != nil {
t.Fatal(err)
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerInline)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "mysource" ||
splunkLoggerDriver.nullMessage.SourceType != "mysourcetype" ||
splunkLoggerDriver.nullMessage.Index != "myindex" ||
!splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.gzipCompressionLevel != gzip.DefaultCompression ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize {
t.Fatal("Values do not match configuration.")
}
messageTime := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("1"), Source: "stdout", Timestamp: messageTime}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 1 {
t.Fatal("Expected one message")
}
if !*hec.gzipEnabled {
t.Fatal("Gzip should be used")
}
message := hec.messages[0]
if message.Time != fmt.Sprintf("%f", float64(messageTime.UnixNano())/float64(time.Second)) ||
message.Host != hostname ||
message.Source != "mysource" ||
message.SourceType != "mysourcetype" ||
message.Index != "myindex" {
t.Fatalf("Unexpected values of message %v", message)
}
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != "1" ||
event["source"] != "stdout" ||
event["tag"] != "container_image_name/container_name" ||
event["attrs"].(map[string]interface{})["a"] != "b" ||
event["attrs"].(map[string]interface{})["foo_finder"] != "bar" ||
len(event) != 4 {
t.Fatalf("Unexpected event in message %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify JSON format
func TestJsonFormat(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkFormatKey: splunkFormatJSON,
splunkGzipCompressionKey: "true",
splunkGzipCompressionLevelKey: "1",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
hostname, err := info.Hostname()
if err != nil {
t.Fatal(err)
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerJSON)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "" ||
splunkLoggerDriver.nullMessage.SourceType != "" ||
splunkLoggerDriver.nullMessage.Index != "" ||
!splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.gzipCompressionLevel != gzip.BestSpeed ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize {
t.Fatal("Values do not match configuration.")
}
message1Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("{\"a\":\"b\"}"), Source: "stdout", Timestamp: message1Time}); err != nil {
t.Fatal(err)
}
message2Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("notjson"), Source: "stdout", Timestamp: message2Time}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 2 {
t.Fatal("Expected two messages")
}
message1 := hec.messages[0]
if message1.Time != fmt.Sprintf("%f", float64(message1Time.UnixNano())/float64(time.Second)) ||
message1.Host != hostname ||
message1.Source != "" ||
message1.SourceType != "" ||
message1.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message1)
}
if event, err := message1.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"].(map[string]interface{})["a"] != "b" ||
event["source"] != "stdout" ||
event["tag"] != "containeriid" ||
len(event) != 3 {
t.Fatalf("Unexpected event in message 1 %v", event)
}
}
message2 := hec.messages[1]
if message2.Time != fmt.Sprintf("%f", float64(message2Time.UnixNano())/float64(time.Second)) ||
message2.Host != hostname ||
message2.Source != "" ||
message2.SourceType != "" ||
message2.Index != "" {
t.Fatalf("Unexpected values of message 2 %v", message2)
}
// If message cannot be parsed as JSON - it should be sent as a line
if event, err := message2.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != "notjson" ||
event["source"] != "stdout" ||
event["tag"] != "containeriid" ||
len(event) != 3 {
t.Fatalf("Unexpected event in message 2 %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify raw format
func TestRawFormat(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkFormatKey: splunkFormatRaw,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
hostname, err := info.Hostname()
assert.NilError(t, err)
loggerDriver, err := New(info)
assert.NilError(t, err)
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerRaw)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "" ||
splunkLoggerDriver.nullMessage.SourceType != "" ||
splunkLoggerDriver.nullMessage.Index != "" ||
splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize ||
string(splunkLoggerDriver.prefix) != "containeriid " {
t.Fatal("Values do not match configuration.")
}
message1Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("{\"a\":\"b\"}"), Source: "stdout", Timestamp: message1Time}); err != nil {
t.Fatal(err)
}
message2Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("notjson"), Source: "stdout", Timestamp: message2Time}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 2 {
t.Fatal("Expected two messages")
}
message1 := hec.messages[0]
if message1.Time != fmt.Sprintf("%f", float64(message1Time.UnixNano())/float64(time.Second)) ||
message1.Host != hostname ||
message1.Source != "" ||
message1.SourceType != "" ||
message1.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message1)
}
if event, err := message1.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "containeriid {\"a\":\"b\"}" {
t.Fatalf("Unexpected event in message 1 %v", event)
}
}
message2 := hec.messages[1]
if message2.Time != fmt.Sprintf("%f", float64(message2Time.UnixNano())/float64(time.Second)) ||
message2.Host != hostname ||
message2.Source != "" ||
message2.SourceType != "" ||
message2.Index != "" {
t.Fatalf("Unexpected values of message 2 %v", message2)
}
if event, err := message2.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "containeriid notjson" {
t.Fatalf("Unexpected event in message 1 %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify raw format with labels
func TestRawFormatWithLabels(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkFormatKey: splunkFormatRaw,
labelsKey: "a",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
ContainerLabels: map[string]string{
"a": "b",
},
}
hostname, err := info.Hostname()
if err != nil {
t.Fatal(err)
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerRaw)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "" ||
splunkLoggerDriver.nullMessage.SourceType != "" ||
splunkLoggerDriver.nullMessage.Index != "" ||
splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize ||
string(splunkLoggerDriver.prefix) != "containeriid a=b " {
t.Fatal("Values do not match configuration.")
}
message1Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("{\"a\":\"b\"}"), Source: "stdout", Timestamp: message1Time}); err != nil {
t.Fatal(err)
}
message2Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("notjson"), Source: "stdout", Timestamp: message2Time}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 2 {
t.Fatal("Expected two messages")
}
message1 := hec.messages[0]
if message1.Time != fmt.Sprintf("%f", float64(message1Time.UnixNano())/float64(time.Second)) ||
message1.Host != hostname ||
message1.Source != "" ||
message1.SourceType != "" ||
message1.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message1)
}
if event, err := message1.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "containeriid a=b {\"a\":\"b\"}" {
t.Fatalf("Unexpected event in message 1 %v", event)
}
}
message2 := hec.messages[1]
if message2.Time != fmt.Sprintf("%f", float64(message2Time.UnixNano())/float64(time.Second)) ||
message2.Host != hostname ||
message2.Source != "" ||
message2.SourceType != "" ||
message2.Index != "" {
t.Fatalf("Unexpected values of message 2 %v", message2)
}
if event, err := message2.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "containeriid a=b notjson" {
t.Fatalf("Unexpected event in message 2 %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify that Splunk Logging Driver can accept tag="" which will allow to send raw messages
// in the same way we get them in stdout/stderr
func TestRawFormatWithoutTag(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkFormatKey: splunkFormatRaw,
tagKey: "",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
hostname, err := info.Hostname()
if err != nil {
t.Fatal(err)
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if !hec.connectionVerified {
t.Fatal("By default connection should be verified")
}
splunkLoggerDriver, ok := loggerDriver.(*splunkLoggerRaw)
if !ok {
t.Fatal("Unexpected Splunk Logging Driver type")
}
if splunkLoggerDriver.url != hec.URL()+"/services/collector/event/1.0" ||
splunkLoggerDriver.auth != "Splunk "+hec.token ||
splunkLoggerDriver.nullMessage.Host != hostname ||
splunkLoggerDriver.nullMessage.Source != "" ||
splunkLoggerDriver.nullMessage.SourceType != "" ||
splunkLoggerDriver.nullMessage.Index != "" ||
splunkLoggerDriver.gzipCompression ||
splunkLoggerDriver.postMessagesFrequency != defaultPostMessagesFrequency ||
splunkLoggerDriver.postMessagesBatchSize != defaultPostMessagesBatchSize ||
splunkLoggerDriver.bufferMaximum != defaultBufferMaximum ||
cap(splunkLoggerDriver.stream) != defaultStreamChannelSize ||
string(splunkLoggerDriver.prefix) != "" {
t.Log(string(splunkLoggerDriver.prefix) + "a")
t.Fatal("Values do not match configuration.")
}
message1Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("{\"a\":\"b\"}"), Source: "stdout", Timestamp: message1Time}); err != nil {
t.Fatal(err)
}
message2Time := time.Now()
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("notjson"), Source: "stdout", Timestamp: message2Time}); err != nil {
t.Fatal(err)
}
message3Time := time.Now()
if err := loggerDriver.Log(&logger.Message{Line: []byte(" "), Source: "stdout", Timestamp: message3Time}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
// message3 would have an empty or whitespace only string in the "event" field
// both of which are not acceptable to HEC
// thus here we must expect 2 messages, not 3
if len(hec.messages) != 2 {
t.Fatal("Expected two messages")
}
message1 := hec.messages[0]
if message1.Time != fmt.Sprintf("%f", float64(message1Time.UnixNano())/float64(time.Second)) ||
message1.Host != hostname ||
message1.Source != "" ||
message1.SourceType != "" ||
message1.Index != "" {
t.Fatalf("Unexpected values of message 1 %v", message1)
}
if event, err := message1.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "{\"a\":\"b\"}" {
t.Fatalf("Unexpected event in message 1 %v", event)
}
}
message2 := hec.messages[1]
if message2.Time != fmt.Sprintf("%f", float64(message2Time.UnixNano())/float64(time.Second)) ||
message2.Host != hostname ||
message2.Source != "" ||
message2.SourceType != "" ||
message2.Index != "" {
t.Fatalf("Unexpected values of message 2 %v", message2)
}
if event, err := message2.EventAsString(); err != nil {
t.Fatal(err)
} else {
if event != "notjson" {
t.Fatalf("Unexpected event in message 2 %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify that we will send messages in batches with default batching parameters,
// but change frequency to be sure that numOfRequests will match expected 17 requests
func TestBatching(t *testing.T) {
if err := os.Setenv(envVarPostMessagesFrequency, "10h"); err != nil {
t.Fatal(err)
}
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
for i := 0; i < defaultStreamChannelSize*4; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != defaultStreamChannelSize*4 {
t.Fatal("Not all messages delivered")
}
for i, message := range hec.messages {
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != fmt.Sprintf("%d", i) {
t.Fatalf("Unexpected event in message %v", event)
}
}
}
// 1 to verify connection and 16 batches
if hec.numOfRequests != 17 {
t.Fatalf("Unexpected number of requests %d", hec.numOfRequests)
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesFrequency, ""); err != nil {
t.Fatal(err)
}
}
// Verify that test is using time to fire events not rare than specified frequency
func TestFrequency(t *testing.T) {
if err := os.Setenv(envVarPostMessagesFrequency, "5ms"); err != nil {
t.Fatal(err)
}
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
for i := 0; i < 10; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
time.Sleep(15 * time.Millisecond)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 10 {
t.Fatal("Not all messages delivered")
}
for i, message := range hec.messages {
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != fmt.Sprintf("%d", i) {
t.Fatalf("Unexpected event in message %v", event)
}
}
}
// 1 to verify connection and 10 to verify that we have sent messages with required frequency,
// but because frequency is too small (to keep test quick), instead of 11, use 9 if context switches will be slow
expectedRequests := 9
if runtime.GOOS == "windows" {
// sometimes in Windows, this test fails with number of requests showing 8. So be more conservative.
expectedRequests = 7
}
if hec.numOfRequests < expectedRequests {
t.Fatalf("Unexpected number of requests %d", hec.numOfRequests)
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesFrequency, ""); err != nil {
t.Fatal(err)
}
}
// Simulate behavior similar to first version of Splunk Logging Driver, when we were sending one message
// per request
func TestOneMessagePerRequest(t *testing.T) {
if err := os.Setenv(envVarPostMessagesFrequency, "10h"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesBatchSize, "1"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, "1"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, "0"); err != nil {
t.Fatal(err)
}
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
for i := 0; i < 10; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 10 {
t.Fatal("Not all messages delivered")
}
for i, message := range hec.messages {
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != fmt.Sprintf("%d", i) {
t.Fatalf("Unexpected event in message %v", event)
}
}
}
// 1 to verify connection and 10 messages
if hec.numOfRequests != 11 {
t.Fatalf("Unexpected number of requests %d", hec.numOfRequests)
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesFrequency, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesBatchSize, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, ""); err != nil {
t.Fatal(err)
}
}
// Driver should not be created when HEC is unresponsive
func TestVerify(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
hec.simulateServerError = true
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
_, err := New(info)
if err == nil {
t.Fatal("Expecting driver to fail, when server is unresponsive")
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify that user can specify to skip verification that Splunk HEC is working.
// Also in this test we verify retry logic.
func TestSkipVerify(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
hec.simulateServerError = true
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkVerifyConnectionKey: "false",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if hec.connectionVerified {
t.Fatal("Connection should not be verified")
}
for i := 0; i < defaultStreamChannelSize*2; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
if len(hec.messages) != 0 {
t.Fatal("No messages should be accepted at this point")
}
hec.simulateErr(false)
for i := defaultStreamChannelSize * 2; i < defaultStreamChannelSize*4; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != defaultStreamChannelSize*4 {
t.Fatal("Not all messages delivered")
}
for i, message := range hec.messages {
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != fmt.Sprintf("%d", i) {
t.Fatalf("Unexpected event in message %v", event)
}
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
// Verify logic for when we filled whole buffer
func TestBufferMaximum(t *testing.T) {
if err := os.Setenv(envVarPostMessagesBatchSize, "2"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, "10"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, "0"); err != nil {
t.Fatal(err)
}
hec := NewHTTPEventCollectorMock(t)
hec.simulateErr(true)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkVerifyConnectionKey: "false",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if hec.connectionVerified {
t.Fatal("Connection should not be verified")
}
for i := 0; i < 11; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
if len(hec.messages) != 0 {
t.Fatal("No messages should be accepted at this point")
}
hec.simulateServerError = false
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 9 {
t.Fatalf("Expected # of messages %d, got %d", 9, len(hec.messages))
}
// First 1000 messages are written to daemon log when buffer was full
for i, message := range hec.messages {
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != fmt.Sprintf("%d", i+2) {
t.Fatalf("Unexpected event in message %v", event)
}
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesBatchSize, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, ""); err != nil {
t.Fatal(err)
}
}
// Verify that we are not blocking close when HEC is down for the whole time
func TestServerAlwaysDown(t *testing.T) {
if err := os.Setenv(envVarPostMessagesBatchSize, "2"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, "4"); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, "0"); err != nil {
t.Fatal(err)
}
hec := NewHTTPEventCollectorMock(t)
hec.simulateServerError = true
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
splunkVerifyConnectionKey: "false",
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
if hec.connectionVerified {
t.Fatal("Connection should not be verified")
}
for i := 0; i < 5; i++ {
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte(fmt.Sprintf("%d", i)), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
if len(hec.messages) != 0 {
t.Fatal("No messages should be sent")
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarPostMessagesBatchSize, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarBufferMaximum, ""); err != nil {
t.Fatal(err)
}
if err := os.Setenv(envVarStreamChannelSize, ""); err != nil {
t.Fatal(err)
}
}
// Cannot send messages after we close driver
func TestCannotSendAfterClose(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
loggerDriver, err := New(info)
if err != nil {
t.Fatal(err)
}
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("message1"), Source: "stdout", Timestamp: time.Now()}); err != nil {
t.Fatal(err)
}
err = loggerDriver.Close()
if err != nil {
t.Fatal(err)
}
Implement optional ring buffer for container logs This allows the user to set a logging mode to "blocking" (default), or "non-blocking", which uses the ring buffer as a proxy to the real log driver. This allows a container to never be blocked on stdio at the cost of dropping log messages. Introduces 2 new log-opts that works for all drivers, `log-mode` and `log-size`. `log-mode` takes a value of "blocking", or "non-blocking" I chose not to implement this as a bool since it is difficult to determine if the mode was set to false vs just not set... especially difficult when merging the default daemon config with the container config. `log-size` takes a size string, e.g. `2MB`, which sets the max size of the ring buffer. When the max size is reached, it will start dropping log messages. ``` BenchmarkRingLoggerThroughputNoReceiver-8 2000000000 36.2 ns/op 856.35 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputWithReceiverDelay0-8 300000000 156 ns/op 198.48 MB/s 32 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay1-8 2000000000 36.1 ns/op 857.80 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay10-8 1000000000 36.2 ns/op 856.53 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay50-8 2000000000 34.7 ns/op 894.65 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay100-8 2000000000 35.1 ns/op 883.91 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay300-8 1000000000 35.9 ns/op 863.90 MB/s 0 B/op 0 allocs/op BenchmarkRingLoggerThroughputConsumeDelay500-8 2000000000 35.8 ns/op 866.88 MB/s 0 B/op 0 allocs/op ``` Signed-off-by: Brian Goff <cpuguy83@gmail.com>
2016-11-22 21:55:27 -05:00
if err := loggerDriver.Log(&logger.Message{Line: []byte("message2"), Source: "stdout", Timestamp: time.Now()}); err == nil {
t.Fatal("Driver should not allow to send messages after close")
}
if len(hec.messages) != 1 {
t.Fatal("Only one message should be sent")
}
message := hec.messages[0]
if event, err := message.EventAsMap(); err != nil {
t.Fatal(err)
} else {
if event["line"] != "message1" {
t.Fatalf("Unexpected event in message %v", event)
}
}
err = hec.Close()
if err != nil {
t.Fatal(err)
}
}
func TestDeadlockOnBlockedEndpoint(t *testing.T) {
hec := NewHTTPEventCollectorMock(t)
go hec.Serve()
info := logger.Info{
Config: map[string]string{
splunkURLKey: hec.URL(),
splunkTokenKey: hec.token,
},
ContainerID: "containeriid",
ContainerName: "/container_name",
ContainerImageID: "contaimageid",
ContainerImageName: "container_image_name",
}
l, err := New(info)
if err != nil {
t.Fatal(err)
}
ctx, unblock := context.WithCancel(context.Background())
hec.withBlock(ctx)
defer unblock()
batchSendTimeout = 1 * time.Second
if err := l.Log(&logger.Message{}); err != nil {
t.Fatal(err)
}
done := make(chan struct{})
go func() {
l.Close()
close(done)
}()
select {
case <-time.After(60 * time.Second):
buf := make([]byte, 1e6)
buf = buf[:runtime.Stack(buf, true)]
t.Logf("STACK DUMP: \n\n%s\n\n", string(buf))
t.Fatal("timeout waiting for close to finish")
case <-done:
}
}