Moving diagnostics into a service (#14832)
* Moving diagnostics into a service * Fixing golint checks * Fixing tests * Renaming from diagnostics to telemetry * Adding missing files * Initializing telemetry earlier in the server startup * Fixing tests * Adding a log for the telemetryID initialization error * Addressing PR review comments * Fixing merge problem * Removing some extra Diagnostics mentions * Making tests pass
Этот коммит содержится в:
Jesús Espino
коммит произвёл
GitHub
GitHub
родитель
f0eb67fa0d
Коммит
44079785eb
124
app/server.go
124
app/server.go
@@ -26,7 +26,6 @@ import (
|
||||
"github.com/gorilla/mux"
|
||||
"github.com/pkg/errors"
|
||||
"github.com/rs/cors"
|
||||
rudder "github.com/rudderlabs/analytics-go"
|
||||
|
||||
"golang.org/x/crypto/acme/autocert"
|
||||
|
||||
@@ -44,6 +43,7 @@ import (
|
||||
"github.com/mattermost/mattermost-server/v5/services/mailservice"
|
||||
"github.com/mattermost/mattermost-server/v5/services/searchengine"
|
||||
"github.com/mattermost/mattermost-server/v5/services/searchengine/bleveengine"
|
||||
"github.com/mattermost/mattermost-server/v5/services/telemetry"
|
||||
"github.com/mattermost/mattermost-server/v5/services/timezones"
|
||||
"github.com/mattermost/mattermost-server/v5/services/tracing"
|
||||
"github.com/mattermost/mattermost-server/v5/services/upgrader"
|
||||
@@ -58,6 +58,9 @@ import (
|
||||
|
||||
var MaxNotificationsPerChannelDefault int64 = 1000000
|
||||
|
||||
// declaring this as var to allow overriding in tests
|
||||
var SENTRY_DSN = "placeholder_sentry_dsn"
|
||||
|
||||
type Server struct {
|
||||
sqlStore *sqlstore.SqlSupplier
|
||||
Store store.Store
|
||||
@@ -136,8 +139,7 @@ type Server struct {
|
||||
clientConfigHash atomic.Value
|
||||
limitedClientConfig atomic.Value
|
||||
|
||||
diagnosticId string
|
||||
rudderClient rudder.Client
|
||||
telemetryService *telemetry.TelemetryService
|
||||
|
||||
phase2PermissionsMigrationComplete bool
|
||||
|
||||
@@ -167,8 +169,7 @@ type Server struct {
|
||||
|
||||
CacheProvider cache.Provider
|
||||
|
||||
tracer *tracing.Tracer
|
||||
timestampLastDiagnosticSent time.Time
|
||||
tracer *tracing.Tracer
|
||||
}
|
||||
|
||||
func NewServer(options ...Option) (*Server, error) {
|
||||
@@ -331,6 +332,8 @@ func NewServer(options ...Option) (*Server, error) {
|
||||
|
||||
s.Store = s.newStore()
|
||||
|
||||
s.telemetryService = telemetry.New(s, s.Store, s.SearchEngine, s.Log)
|
||||
|
||||
emailService, err := NewEmailService(s)
|
||||
if err != nil {
|
||||
return nil, errors.Wrapf(err, "unable to initialize email service")
|
||||
@@ -363,7 +366,6 @@ func NewServer(options ...Option) (*Server, error) {
|
||||
return nil, errors.Wrapf(err, "unable to ensure first run timestamp")
|
||||
}
|
||||
|
||||
s.ensureDiagnosticId()
|
||||
s.regenerateClientConfig()
|
||||
|
||||
s.clusterLeaderListenerId = s.AddClusterLeaderChangedListener(func() {
|
||||
@@ -519,7 +521,13 @@ func (s *Server) RunJobs() {
|
||||
runSecurityJob(s)
|
||||
})
|
||||
s.Go(func() {
|
||||
runDiagnosticsJob(s)
|
||||
firstRun, err := s.getFirstServerRunTimestamp()
|
||||
if err != nil {
|
||||
mlog.Warn("Fetching time of first server run failed. Setting to 'now'.")
|
||||
s.ensureFirstServerRunTimestamp()
|
||||
firstRun = utils.MillisFromTime(time.Now())
|
||||
}
|
||||
s.telemetryService.RunTelemetryJob(firstRun)
|
||||
})
|
||||
s.Go(func() {
|
||||
runSessionCleanupJob(s)
|
||||
@@ -676,9 +684,9 @@ func (s *Server) Shutdown() error {
|
||||
}
|
||||
}
|
||||
|
||||
err := s.shutdownDiagnostics()
|
||||
err := s.telemetryService.Shutdown()
|
||||
if err != nil {
|
||||
mlog.Error("Unable to cleanly shutdown diagnostic client", mlog.Err(err))
|
||||
mlog.Error("Unable to cleanly shutdown telemetry client", mlog.Err(err))
|
||||
}
|
||||
|
||||
s.StopHTTPServer()
|
||||
@@ -1105,34 +1113,6 @@ func runSecurityJob(s *Server) {
|
||||
}, time.Hour*4)
|
||||
}
|
||||
|
||||
func doDiagnosticsIfNeeded(s *Server, firstRun time.Time) {
|
||||
hoursSinceFirstServerRun := time.Since(firstRun).Hours()
|
||||
// Send once every 10 minutes for the first hour
|
||||
// Send once every hour thereafter for the first 12 hours
|
||||
// Send at the 24 hour mark and every 24 hours after
|
||||
if hoursSinceFirstServerRun < 1 {
|
||||
doDiagnostics(s)
|
||||
} else if hoursSinceFirstServerRun <= 12 && time.Since(s.timestampLastDiagnosticSent) >= time.Hour {
|
||||
doDiagnostics(s)
|
||||
} else if hoursSinceFirstServerRun > 12 && time.Since(s.timestampLastDiagnosticSent) >= 24*time.Hour {
|
||||
doDiagnostics(s)
|
||||
}
|
||||
}
|
||||
|
||||
func runDiagnosticsJob(s *Server) {
|
||||
// Send on boot
|
||||
doDiagnostics(s)
|
||||
firstRun, err := s.getFirstServerRunTimestamp()
|
||||
if err != nil {
|
||||
mlog.Warn("Fetching time of first server run failed. Setting to 'now'.")
|
||||
s.ensureFirstServerRunTimestamp()
|
||||
firstRun = utils.MillisFromTime(time.Now())
|
||||
}
|
||||
model.CreateRecurringTask("Diagnostics", func() {
|
||||
doDiagnosticsIfNeeded(s, utils.TimeFromMillis(firstRun))
|
||||
}, time.Minute*10)
|
||||
}
|
||||
|
||||
func runTokenCleanupJob(s *Server) {
|
||||
doTokenCleanup(s)
|
||||
model.CreateRecurringTask("Token Cleanup", func() {
|
||||
@@ -1172,13 +1152,6 @@ func doSecurity(s *Server) {
|
||||
s.DoSecurityUpdateCheck()
|
||||
}
|
||||
|
||||
func doDiagnostics(s *Server) {
|
||||
if *s.Config().LogSettings.EnableDiagnostics {
|
||||
s.timestampLastDiagnosticSent = time.Now()
|
||||
s.SendDailyDiagnostics()
|
||||
}
|
||||
}
|
||||
|
||||
func doTokenCleanup(s *Server) {
|
||||
s.Store.Token().Cleanup()
|
||||
}
|
||||
@@ -1368,39 +1341,6 @@ func (s *Server) stopSearchEngine() {
|
||||
}
|
||||
}
|
||||
|
||||
// initDiagnostics initialises the Rudder client for the diagnostics system.
|
||||
func (s *Server) initDiagnostics(endpoint string, rudderKey string) {
|
||||
if s.rudderClient == nil {
|
||||
config := rudder.Config{}
|
||||
config.Logger = rudder.StdLogger(s.Log.StdLog(mlog.String("source", "rudder")))
|
||||
config.Endpoint = endpoint
|
||||
// For testing
|
||||
if endpoint != RUDDER_DATAPLANE_URL {
|
||||
config.Verbose = true
|
||||
config.BatchSize = 1
|
||||
}
|
||||
client, err := rudder.NewWithConfig(rudderKey, endpoint, config)
|
||||
if err != nil {
|
||||
mlog.Error("Failed to create Rudder instance", mlog.Err(err))
|
||||
return
|
||||
}
|
||||
client.Enqueue(rudder.Identify{
|
||||
UserId: s.diagnosticId,
|
||||
})
|
||||
|
||||
s.rudderClient = client
|
||||
}
|
||||
}
|
||||
|
||||
// shutdownDiagnostics closes the diagnostics system Rudder client.
|
||||
func (s *Server) shutdownDiagnostics() error {
|
||||
if s.rudderClient != nil {
|
||||
return s.rudderClient.Close()
|
||||
}
|
||||
|
||||
return nil
|
||||
}
|
||||
|
||||
func (s *Server) FileBackend() (filesstore.FileBackend, *model.AppError) {
|
||||
license := s.License()
|
||||
return filesstore.NewFileBackend(&s.Config().FileSettings, license != nil && *license.Features.Compliance)
|
||||
@@ -1421,25 +1361,6 @@ func (s *Server) ClusterHealthScore() int {
|
||||
return s.Cluster.HealthScore()
|
||||
}
|
||||
|
||||
func (s *Server) ensureDiagnosticId() {
|
||||
if s.diagnosticId != "" {
|
||||
return
|
||||
}
|
||||
props, err := s.Store.System().Get()
|
||||
if err != nil {
|
||||
return
|
||||
}
|
||||
|
||||
id := props[model.SYSTEM_DIAGNOSTIC_ID]
|
||||
if len(id) == 0 {
|
||||
id = model.NewId()
|
||||
systemID := &model.System{Name: model.SYSTEM_DIAGNOSTIC_ID, Value: id}
|
||||
s.Store.System().Save(systemID)
|
||||
}
|
||||
|
||||
s.diagnosticId = id
|
||||
}
|
||||
|
||||
func (s *Server) configOrLicenseListener() {
|
||||
s.regenerateClientConfig()
|
||||
}
|
||||
@@ -1469,3 +1390,14 @@ func (s *Server) initJobs() {
|
||||
s.Jobs.Migrations = jobsMigrationsInterface(s)
|
||||
}
|
||||
}
|
||||
|
||||
func (s *Server) TelemetryId() string {
|
||||
if s.telemetryService == nil {
|
||||
return ""
|
||||
}
|
||||
return s.telemetryService.TelemetryID
|
||||
}
|
||||
|
||||
func (s *Server) HttpService() httpservice.HTTPService {
|
||||
return s.HTTPService
|
||||
}
|
||||
|
||||
Ссылка в новой задаче
Block a user