report GC stats even in the event of a GC worker error

When customers seed their SpiceDB clusters and
use idempotent TOUCH operations, large amounts of garbage
may be generated that all will expire around the same time.

This makes the GC workers take longer to process, to the point
they may exceed the deadline. In these scenarios the code is
not reporting metrics, which hides how much garbage was being collected.

This changes the code so that GC metrics are issued, even in the event
of an error.

internal/datastore/common/gc.go

@@ -184,22 +184,26 @@ func RunGarbageCollection(gc GarbageCollector, window, timeout time.Duration) er
 	}
 
 	collected, err := gc.DeleteBeforeTx(ctx, watermark)
+
+	// even if an error happened, garbage would have been collected. This makes sure these are reflected even if the
+	// worker eventually fails or times out.
+	gcRelationshipsCounter.Add(float64(collected.Relationships))
+	gcTransactionsCounter.Add(float64(collected.Transactions))
+	gcNamespacesCounter.Add(float64(collected.Namespaces))
+	collectionDuration := time.Since(startTime)
+	gcDurationHistogram.Observe(collectionDuration.Seconds())
+
 	if err != nil {
 		return fmt.Errorf("error deleting in gc: %w", err)
 	}
 
-	collectionDuration := time.Since(startTime)
 	log.Ctx(ctx).Info().
 		Stringer("highestTxID", watermark).
 		Dur("duration", collectionDuration).
 		Time("nowTime", now).
 		Interface("collected", collected).
 		Msg("datastore garbage collection completed successfully")
 
-	gcDurationHistogram.Observe(collectionDuration.Seconds())
-	gcRelationshipsCounter.Add(float64(collected.Relationships))
-	gcTransactionsCounter.Add(float64(collected.Transactions))
-	gcNamespacesCounter.Add(float64(collected.Namespaces))
 	gc.MarkGCCompleted()
 	return nil
 }