Merge pull request #2153 from binwiederhier/cluster-status-self

ntfy cluster status: this node as the table's first row

cmd/cluster.go

@@ -132,20 +132,17 @@ func clusterStatusBaseURL(listen string) (string, error) {
 }
 
 func printClusterStatus(w io.Writer, status *cluster.Status) {
-	leader, healthy := "not leader", "healthy"
+	// This node first, then its peers; the link columns are the peers' only
+	self, live := string(status.NodeID)+" (this node)", "yes"
 	if status.Leader {
-		leader = "leader"
+		self = string(status.NodeID) + " (this node, leader)"
 	}
 	if !status.Healthy {
-		healthy = "NOT healthy (registry heartbeat stale)"
-	}
-	fmt.Fprintf(w, "Node %s: %s, %s, %s\n\n", status.NodeID, status.Role, leader, healthy)
-	if len(status.Peers) == 0 {
-		fmt.Fprintln(w, "No peers.")
-		return
+		live = "no"
 	}
 	tw := tabwriter.NewWriter(w, 0, 0, 2, ' ', 0)
-	fmt.Fprintln(tw, "PEER\tROLE\tLIVE\tHEARTBEAT\tSENT OK\tFAILED\tQUEUED\tDROPPED\tLAST SENT\tRECEIVED\tLAST RECEIVED")
+	fmt.Fprintln(tw, "NODE\tROLE\tLIVE\tHEARTBEAT\tSENT OK\tFAILED\tQUEUED\tDROPPED\tLAST SENT\tRECEIVED\tLAST RECEIVED")
+	fmt.Fprintf(tw, "%s\t%s\t%s\t-\t-\t-\t-\t-\t-\t-\t-\n", self, status.Role, live)
 	for _, p := range status.Peers {
 		live, heartbeat := "no", "last seen "+clusterStatusAge(status.Time, p.LastSeen)
 		if p.Live {
@@ -156,6 +153,9 @@ func printClusterStatus(w io.Writer, status *cluster.Status) {
 	}
 	tw.Flush()
 	problems := clusterStatusProblems(status, "")
+	if !status.Healthy {
+		problems = append([]string{fmt.Sprintf("%s (this node): not healthy, its registry heartbeat is stale", status.NodeID)}, problems...)
+	}
 	if len(problems) == 0 {
 		return
 	}

cmd/cluster_test.go

@@ -84,6 +84,21 @@ func TestCLI_Cluster_Status(t *testing.T) {
 	require.Contains(t, out, "nyc3-app-1: no batch accepted") // The warning that matters
 }
 
+func TestCLI_Cluster_Status_ThisNodeInTheList(t *testing.T) {
+	// The node itself is the table's first row, not a line above it; a leader and an unhealthy node say so
+	status := &cluster.Status{NodeID: "nyc3-app-1", Role: cluster.RoleActive, Leader: true, Healthy: false, Time: 1792330000, Peers: []*cluster.PeerStatus{{
+		NodeID: "nyc3-app-2", Role: cluster.RolePassive, Live: true, HeartbeatAge: 1, Out: cluster.PeerOutbound{BatchesOK: 7, LastOK: 1792329999},
+	}}}
+	addr := newTestClusterStatusServerFor(t, func() *cluster.Status { return status })
+	app, _, stdout, _ := newTestApp()
+	require.Nil(t, app.Run([]string{"ntfy", "cluster", "--experimental-cluster-listen", addr, "--experimental-cluster-secret", "s3cret", "status"}))
+	lines := strings.Split(strings.TrimSpace(stdout.String()), "\n")
+	require.True(t, strings.HasPrefix(lines[0], "NODE "), "the table comes first:\n%s", stdout.String())
+	require.Regexp(t, `^nyc3-app-1 \(this node, leader\)\s+active\s+no\s`, lines[1])
+	require.Regexp(t, `^nyc3-app-2\s+passive\s+yes\s+1s ago\s+7\s`, lines[2])
+	require.Contains(t, stdout.String(), "nyc3-app-1 (this node): not healthy, its registry heartbeat is stale")
+}
+
 func TestCLI_Cluster_Status_JSON(t *testing.T) {
 	addr := newTestClusterStatusServer(t)
 	app, _, stdout, _ := newTestApp()

docs/config.md

@@ -2376,15 +2376,14 @@ once.
 `/v1/health` only covers a node's own database heartbeat. A link between two nodes that works in one direction only,
 e.g. because a firewall drops the cluster port one way, keeps both nodes healthy while messages published on one of them
 reach the other only late, through the database. To see each node's view of its peers, run `ntfy cluster status` on the
-node. It reads `experimental-cluster-listen` and `experimental-cluster-secret` from the server config, and shows per
-peer the message batches sent there and received from there:
+node. It reads `experimental-cluster-listen` and `experimental-cluster-secret` from the server config, and lists the
+node itself (its role, and whether it leads), then per peer the message batches sent there and received from there:
 
 ```
 $ ntfy cluster status
-Node nyc3-app-2: passive, not leader, healthy
-
-PEER        ROLE    LIVE  HEARTBEAT  SENT OK  FAILED  QUEUED  DROPPED  LAST SENT  RECEIVED  LAST RECEIVED
-nyc3-app-1  active  yes   2s ago     0        219     0       0        never      5120      1s ago
+NODE                    ROLE     LIVE  HEARTBEAT  SENT OK  FAILED  QUEUED  DROPPED  LAST SENT  RECEIVED  LAST RECEIVED
+nyc3-app-2 (this node)  passive  yes   -          -        -       -       -        -          -         -
+nyc3-app-1              active   yes   2s ago     0        219     0       0        never      5120      1s ago
 
 Problems:
   nyc3-app-1: no batch accepted, 219 failed; last error 10s ago: Post "http://10.108.0.2:2587/v1/cluster/message": context deadline exceeded

docs/releases.md

@@ -2356,6 +2356,7 @@ and the [ntfy Android app](https://github.com/binwiederhier/ntfy-android/release
 **Bug fixes + maintenance:**
 
 * Do not send TCP keep-alive probes on the client listeners (`listen-http`, `listen-https`): Go sends one every 15 seconds on an idle connection by default, which, without a proxy in front of ntfy, woke phones holding a stream three times as often as ntfy's own keepalive event ([#2152](https://github.com/binwiederhier/ntfy/pull/2152))
+* `ntfy cluster status` lists the node itself as the first row of the table (marked "this node", and "leader" if it leads) instead of a line above it ([#2153](https://github.com/binwiederhier/ntfy/pull/2153))
 
 ### ntfy iOS app v1.8.0 (UNRELEASED)