satellite/metrics: join for monitoring

We don't want metrics to trigger the metaloop on it's own.

Change-Id: I648272a0e3043173b0a377697db1f172712bdda5

private/testplanet/satellite.go

@@ -603,9 +603,7 @@ func (planet *Planet) newSatellite(ctx context.Context, prefix string, index int
 			MaxOrderLimitSendCount:       3,
 			NodeMinAgeInMonths:           0,
 		},
-		Metrics: metrics.Config{
-			ChoreInterval: defaultInterval,
-		},
+		Metrics: metrics.Config{},
 	}
 
 	if planet.config.Reconfigure.Satellite != nil {

satellite/metrics/chore.go

@@ -23,7 +23,6 @@ var (
 
 // Config contains configurable values for metrics collection.
 type Config struct {
-	ChoreInterval time.Duration `help:"the time between each metrics chore run" releaseDefault:"15m" devDefault:"15m"`
 }
 
 // Chore implements the metrics chore.
@@ -40,9 +39,10 @@ type Chore struct {
 // NewChore creates a new instance of the metrics chore.
 func NewChore(log *zap.Logger, config Config, loop *metaloop.Service) *Chore {
 	return &Chore{
-		log:          log,
-		config:       config,
-		Loop:         sync2.NewCycle(config.ChoreInterval),
+		log:    log,
+		config: config,
+		// This chore monitors metainfo loop, so it's fine to use very small cycle time.
+		Loop:         sync2.NewCycle(time.Nanosecond),
 		metainfoLoop: loop,
 	}
 }
@@ -56,7 +56,7 @@ func (chore *Chore) Run(ctx context.Context) (err error) {
 
 		chore.Counter = NewCounter()
 
-		err = chore.metainfoLoop.Join(ctx, chore.Counter)
+		err = chore.metainfoLoop.Monitor(ctx, chore.Counter)
 		if err != nil {
 			chore.log.Error("error joining metainfoloop", zap.Error(err))
 			return nil

scripts/testdata/satellite-config.yaml.lock

@@ -415,9 +415,6 @@ identity.key-path: /root/.local/share/storj/identity/satellite/identity.key
 # application suffix
 # metrics.app-suffix: -release
 
-# the time between each metrics chore run
-# metrics.chore-interval: 15m0s
-
 # instance id prefix
 # metrics.instance-prefix: ""