Always collect metrics about top backend requests

Previously, it was only done in debug mode. This makes some tabs in
`tctl top` empty, when auth server is not in debug mode.

lib/backend/report.go

@@ -33,9 +33,6 @@ import (
 type ReporterConfig struct {
 	// Backend is a backend to wrap
 	Backend Backend
-	// TrackTopRequests turns on tracking of top
-	// requests on
-	TrackTopRequests bool
 	// Component is a component name to report
 	Component string
 }
@@ -221,9 +218,6 @@ func (s *Reporter) Migrate(ctx context.Context) error { return s.Backend.Migrate
 
 // trackRequests tracks top requests, endKey is supplied for ranges
 func (s *Reporter) trackRequest(opType OpType, key []byte, endKey []byte) {
-	if !s.TrackTopRequests {
-		return
-	}
 	if len(key) == 0 {
 		return
 	}

lib/service/service.go

@@ -1320,9 +1320,8 @@ func (process *TeleportProcess) newAccessCache(cfg accessCacheConfig) (*cache.Ca
 		cacheBackend = liteBackend
 	}
 	reporter, err := backend.NewReporter(backend.ReporterConfig{
-		Component:        teleport.ComponentCache,
-		Backend:          cacheBackend,
-		TrackTopRequests: process.Config.Debug,
+		Component: teleport.ComponentCache,
+		Backend:   cacheBackend,
 	})
 	if err != nil {
 		return nil, trace.Wrap(err)
@@ -2392,9 +2391,8 @@ func (process *TeleportProcess) initAuthStorage() (bk backend.Backend, err error
 		return nil, trace.Wrap(err)
 	}
 	reporter, err := backend.NewReporter(backend.ReporterConfig{
-		Component:        teleport.ComponentBackend,
-		Backend:          backend.NewSanitizer(bk),
-		TrackTopRequests: process.Config.Debug,
+		Component: teleport.ComponentBackend,
+		Backend:   backend.NewSanitizer(bk),
 	})
 	if err != nil {
 		return nil, trace.Wrap(err)