address irfan nits and lint errors

pkg/cmd/roachprod/flags.go

@@ -244,7 +244,9 @@ func initFlags() {
 		"open", false, "open the grafana dashboard url on the browser")
 
 	grafanaStopCmd.Flags().StringVar(&grafanaDumpDir, "dump-dir", "",
-		"the absolute path to dump prometheus data to. If empty, no data will get dumped")
+		"the absolute path on the machine running roachprod to dump prometheus data to.\n"+
+			"In the dump-dir, the 'prometheus-docker-run.sh' bash script allows the user to interact\n"+
+			"with the dumped data. If empty, no data will get dumped.")
 
 	for _, cmd := range []*cobra.Command{createCmd, destroyCmd, extendCmd, logsCmd} {
 		cmd.Flags().StringVarP(&username, "username", "u", os.Getenv("ROACHPROD_USER"),

pkg/cmd/roachprod/main.go

@@ -886,8 +886,8 @@ var getProvidersCmd = &cobra.Command{
 }
 
 var grafanaStartCmd = &cobra.Command{
-	Use:   `grafana-start <cluster> [<grafanaConfig>]`,
-	Short: `spins up a prometheus and grafana instances on the last node provided`,
+	Use:   `grafana-start <cluster>`,
+	Short: `spins up a prometheus and grafana instances on the last node in the cluster`,
 	Long: `spins up a prometheus and grafana instances on the highest numbered node in the cluster
 and will scrape from all nodes in the cluster`,
 	Args: cobra.ExactArgs(1),
@@ -900,11 +900,10 @@ and will scrape from all nodes in the cluster`,
 var grafanaStopCmd = &cobra.Command{
 	Use:   `grafana-stop <cluster>`,
 	Short: `spins down prometheus and grafana instances on the last node in the cluster`,
-	Long: `spins down the prometheus and grafana instances on provided roachprod node and
-dumps the prometheus data at the root of the last node in the cluster`,
-	Args: cobra.ExactArgs(1),
+	Long:  `spins down the prometheus and grafana instances on the last node in the cluster`,
+	Args:  cobra.ExactArgs(1),
 	Run: wrap(func(cmd *cobra.Command, args []string) error {
-		return roachprod.StopGrafana(context.Background(), roachprodLibraryLogger, args[0], "")
+		return roachprod.StopGrafana(context.Background(), roachprodLibraryLogger, args[0], grafanaDumpDir)
 	}),
 }
 

pkg/cmd/roachtest/tests/canary.go

@@ -109,17 +109,6 @@ var canaryRetryOptions = retry.Options{
 	MaxRetries:     10,
 }
 
-type repeatRunner struct {
-	T test.Test
-	C cluster.Cluster
-}
-
-func (rr repeatRunner) repeatRunE(
-	ctx context.Context, node option.NodeListOption, operation string, args ...string,
-) error {
-	return repeatRunE(ctx, rr.T, rr.C, node, operation, args...)
-}
-
 // repeatRunE is the same function as c.RunE but with an automatic retry loop.
 func repeatRunE(
 	ctx context.Context,

pkg/roachprod/install/cluster_synced.go

@@ -772,7 +772,7 @@ var roachprodRetryOptions = retry.Options{
 	MaxRetries:     10,
 }
 
-// repeatRun is the same function as c.Run but with an automatic retry loop.
+// RepeatRun is the same function as c.Run, but with an automatic retry loop.
 func (c *SyncedCluster) RepeatRun(
 	ctx context.Context, l *logger.Logger, stdout, stderr io.Writer, nodes Nodes, title,
 	cmd string,

pkg/roachprod/prometheus/prometheus.go

@@ -436,7 +436,7 @@ func Shutdown(
 		"stop prometheus",
 		"sudo systemctl stop prometheus || echo 'Stopped prometheus'",
 	); err != nil {
-		l.Printf("Failed to stop prometheus server")
+		l.Printf("Failed to stop prometheus server: %v", err)
 	}
 	return nil
 }

pkg/roachprod/roachprod.go

@@ -1430,6 +1430,7 @@ func StopGrafana(ctx context.Context, l *logger.Logger, clusterName string, dump
 	return nil
 }
 
+// GrafanaURL returns a url to the grafana dashboard
 func GrafanaURL(
 	ctx context.Context, l *logger.Logger, clusterName string, openInBrowser bool,
 ) ([]string, error) {