Merge pull request #358 from jfontan/improvement/add-tracing-flag

cmd/server: add trace parameters to enable jaeger tracing

Author: ajnavarro

Gopkg.lock

@@ -54,3 +54,7 @@
 [[constraint]]
   name = "gopkg.in/bblfsh/sdk.v1"
   version = "1.16.0"
+
+[[constraint]]
+  name = "github.com/uber/jaeger-client-go"
+  version = "^2.7.0"

Gopkg.toml

@@ -10,8 +10,10 @@ import (
 	"github.com/src-d/gitbase/internal/function"
 	"github.com/src-d/gitbase/internal/rule"
 
+	"github.com/opentracing/opentracing-go"
 	gopilosa "github.com/pilosa/go-pilosa"
 	"github.com/sirupsen/logrus"
+	"github.com/uber/jaeger-client-go/config"
 	sqle "gopkg.in/src-d/go-mysql-server.v0"
 	"gopkg.in/src-d/go-mysql-server.v0/server"
 	"gopkg.in/src-d/go-mysql-server.v0/sql/analyzer"
@@ -22,9 +24,10 @@ import (
 const (
 	ServerDescription = "Starts a gitbase server instance"
 	ServerHelp        = ServerDescription + "\n\n" +
-		"By default when gitbase encounters and error in a repository it\n" +
+		"By default when gitbase encounters an error in a repository it\n" +
 		"stops the query. With GITBASE_SKIP_GIT_ERRORS variable it won't\n" +
 		"complain and just skip those rows or repositories."
+	TracerServiceName = "gitbase"
 )
 
 // Server represents the `server` command of gitbase cli tool.
@@ -39,16 +42,24 @@ type Server struct {
 	PilosaURL     string   `long:"pilosa" default:"http://localhost:10101" description:"URL to your pilosa server" env:"PILOSA_ENDPOINT"`
 	IndexDir      string   `short:"i" long:"index" default:"/var/lib/gitbase/index" description:"Directory where the gitbase indexes information will be persisted." env:"GITBASE_INDEX_DIR"`
 	DisableSquash bool     `long:"no-squash" description:"Disables the table squashing."`
-	// IgnoreGitErrors by default when gitbase encounters and error in a
-	// repository it stops the query. With this parameter it won't complain and
-	// just skip those rows or repositories.
+	TraceEnabled  bool     `long:"trace" env:"GITBASE_TRACE" description:"Enables jaeger tracing"`
+
+	// SkipGitErrors disables failing when Git errors are found.
 	SkipGitErrors bool
 
 	engine *sqle.Engine
 	pool   *gitbase.RepositoryPool
 	name   string
 }
 
+type jaegerLogrus struct {
+	*logrus.Entry
+}
+
+func (l *jaegerLogrus) Error(s string) {
+	l.Entry.Error(s)
+}
+
 // Execute starts a new gitbase server based on provided configuration, it
 // honors the go-flags.Commander interface.
 func (c *Server) Execute(args []string) error {
@@ -66,12 +77,43 @@ func (c *Server) Execute(args []string) error {
 		{Password: c.Password},
 	}
 
+	var tracer opentracing.Tracer
+	if c.TraceEnabled {
+		cfg, err := config.FromEnv()
+		if err != nil {
+			logrus.WithField("error", err).
+				Fatal("unable to read jaeger environment")
+			return err
+		}
+
+		if cfg.ServiceName == "" {
+			cfg.ServiceName = TracerServiceName
+		}
+
+		logger := &jaegerLogrus{logrus.WithField("subsystem", "jaeger")}
+
+		t, closer, err := cfg.NewTracer(
+			config.Logger(logger),
+		)
+
+		if err != nil {
+			logrus.WithField("error", err).Fatal("unable to initialize tracer")
+			return err
+		}
+
+		tracer = t
+		defer closer.Close()
+
+		logrus.Info("tracing enabled")
+	}
+
 	hostString := net.JoinHostPort(c.Host, strconv.Itoa(c.Port))
 	s, err := server.NewServer(
 		server.Config{
 			Protocol: "tcp",
 			Address:  hostString,
 			Auth:     auth,
+			Tracer:   tracer,
 		},
 		c.engine,
 		gitbase.NewSessionBuilder(c.pool,

cmd/gitbase/command/server.go

@@ -2,17 +2,40 @@
 
 ## Environment variables
 
-| Name                             | Description                                         |
-|:---------------------------------|:----------------------------------------------------|
-| `BBLFSH_ENDPOINT`                | bblfshd endpoint, default "127.0.0.1:9432"          |
-| `PILOSA_ENDPOINT`                | pilosa endpoint, default "http://localhost:10101"   |
-| `GITBASE_BLOBS_MAX_SIZE`         | maximum blob size to return in MiB, default 5 MiB   |
-| `GITBASE_BLOBS_ALLOW_BINARY`     | enable retrieval of binary blobs, default `false`   |
-| `GITBASE_SKIP_GIT_ERRORS`        | do not stop queries on git errors, default disabled |
+| Name                         | Description                                                 |
+|:-----------------------------|:------------------------------------------------------------|
+| `BBLFSH_ENDPOINT`            | bblfshd endpoint, default "127.0.0.1:9432"                  |
+| `PILOSA_ENDPOINT`            | pilosa endpoint, default "http://localhost:10101"           |
+| `GITBASE_BLOBS_MAX_SIZE`     | maximum blob size to return in MiB, default 5 MiB           |
+| `GITBASE_BLOBS_ALLOW_BINARY` | enable retrieval of binary blobs, default `false`           |
+| `GITBASE_SKIP_GIT_ERRORS`    | do not stop queries on git errors, default disabled         |
+| `GITBASE_INDEX_DIR`          | directory to save indexes, default `/var/lib/gitbase/index` |
+| `GITBASE_TRACE`              | enable jaeger tracing, default disabled                     |
+
+### Jaeger tracing variables
+
+*Extracted from https://github.com/jaegertracing/jaeger-client-go/blob/master/README.md*
+
+Property| Description
+--- | ---
+JAEGER_SERVICE_NAME | The service name
+JAEGER_AGENT_HOST | The hostname for communicating with agent via UDP
+JAEGER_AGENT_PORT | The port for communicating with agent via UDP
+JAEGER_REPORTER_LOG_SPANS | Whether the reporter should also log the spans
+JAEGER_REPORTER_MAX_QUEUE_SIZE | The reporter's maximum queue size
+JAEGER_REPORTER_FLUSH_INTERVAL | The reporter's flush interval (ms)
+JAEGER_SAMPLER_TYPE | The sampler type
+JAEGER_SAMPLER_PARAM | The sampler parameter (number)
+JAEGER_SAMPLER_MANAGER_HOST_PORT | The host name and port when using the remote controlled sampler
+JAEGER_SAMPLER_MAX_OPERATIONS | The maximum number of operations that the sampler will keep track of
+JAEGER_SAMPLER_REFRESH_INTERVAL | How often the remotely controlled sampler will poll jaeger-agent for the appropriate sampling strategy
+JAEGER_TAGS | A comma separated list of `name = value` tracer level tags, which get added to all reported spans. The value can also refer to an environment variable using the format `${envVarName:default}`, where the `:default` is optional, and identifies a value to be used if the environment variable cannot be found
+JAEGER_DISABLED | Whether the tracer is disabled or not. If true, the default `opentracing.NoopTracer` is used.
+JAEGER_RPC_METRICS | Whether to store RPC metrics
 
 ## Command line arguments
 
-```bash
+```
 Please specify one command of: server or version
 Usage:
   gitbase [OPTIONS] <server | version>
@@ -27,7 +50,7 @@ Available commands:
 
 `server` command contains the following options:
 
-```bash
+```
 Usage:
   gitbase [OPTIONS] server [server-OPTIONS]
 
@@ -38,17 +61,22 @@ stops the query. With GITBASE_SKIP_GIT_ERRORS variable it won't
 complain and just skip those rows or repositories.
 
 Help Options:
-  -h, --help          Show this help message
+  -h, --help           Show this help message
 
 [server command options]
-      -v              Activates the verbose mode
-      -g, --git=      Path where the git repositories are located, multiple directories can be defined. Accepts globs.
-          --siva=     Path where the siva repositories are located, multiple directories can be defined. Accepts globs.
-      -h, --host=     Host where the server is going to listen (default: localhost)
-      -p, --port=     Port where the server is going to listen (default: 3306)
-      -u, --user=     User name used for connection (default: root)
-      -P, --password= Password used for connection
-          --pilosa=   URL to your pilosa server (default: http://localhost:10101)
-      -i, --index=    Directory where the gitbase indexes information will be persisted. (default: /var/lib/gitbase/index)
-          --no-squash Disables the table squashing.
+      -v               Activates the verbose mode
+      -g, --git=       Path where the git repositories are located, multiple directories can
+                       be defined. Accepts globs.
+          --siva=      Path where the siva repositories are located, multiple directories can
+                       be defined. Accepts globs.
+      -h, --host=      Host where the server is going to listen (default: localhost)
+      -p, --port=      Port where the server is going to listen (default: 3306)
+      -u, --user=      User name used for connection (default: root)
+      -P, --password=  Password used for connection
+          --pilosa=    URL to your pilosa server (default: http://localhost:10101)
+                       [$PILOSA_ENDPOINT]
+      -i, --index=     Directory where the gitbase indexes information will be persisted.
+                       (default: /var/lib/gitbase/index) [$GITBASE_INDEX_DIR]
+          --no-squash  Disables the table squashing.
+          --trace      Enables jaeger tracing [$GITBASE_TRACE]
 ```