chore(telemetry): dot-namespace grouped telemetry keys (#968)

- Rename grouped `workspace.agent.state_transitioned` properties to
  dot-namespaced `status.*` and `lifecycle_state.*` keys.
- Rename `http.requests` aggregate metric families to `count.*` and
  `duration.*` keys, preserving the snake_case leaves
  (`count.{1xx,2xx,3xx,4xx,5xx,network_error}`, `duration.p{50,95,99}_ms`).
- Apply the optional `workspace.state_transitioned` clarity rename:
  `transition` → `build.transition`, `reason` → `build.reason`.
- Replace old keys only; do not dual-emit legacy keys.

Closes #954

Author: EhabY

src/instrumentation/workspace.ts

@@ -26,8 +26,8 @@ const PROVISIONING_STATUSES: ReadonlySet<WorkspaceStatus> = new Set([
 
 interface ObservedWorkspaceState {
 	readonly status: WorkspaceStatus;
-	readonly transition: WorkspaceBuild["transition"];
-	readonly reason: WorkspaceBuild["reason"];
+	readonly buildTransition: WorkspaceBuild["transition"];
+	readonly buildReason: WorkspaceBuild["reason"];
 	readonly observedAtMs: number;
 }
 
@@ -53,12 +53,16 @@ export class WorkspaceStateTelemetry {
 	) {}
 
 	public observe(workspace: Workspace): void {
-		const { status, transition, reason } = workspace.latest_build;
+		const {
+			status,
+			transition: buildTransition,
+			reason: buildReason,
+		} = workspace.latest_build;
 		const previous = this.observed;
 		if (
 			previous?.status === status &&
-			previous.transition === transition &&
-			previous.reason === reason
+			previous.buildTransition === buildTransition &&
+			previous.buildReason === buildReason
 		) {
 			return;
 		}
@@ -86,20 +90,25 @@ export class WorkspaceStateTelemetry {
 				workspaceName: this.workspaceName,
 				from: previous?.status ?? INITIAL_STATE,
 				to: status,
-				transition,
-				reason,
+				"build.transition": buildTransition,
+				"build.reason": buildReason,
 			},
 			measurements,
 		);
-		this.observed = { status, transition, reason, observedAtMs: now };
+		this.observed = {
+			status,
+			buildTransition,
+			buildReason,
+			observedAtMs: now,
+		};
 	}
 }
 
 /**
  * Emits `workspace.agent.state_transitioned` as the agent's `status` and
  * `lifecycle_state` change. The agent has two state dimensions so the event
- * carries qualified `fromStatus`/`toStatus` and `fromLifecycleState`/
- * `toLifecycleState` properties. Construct one per workspace.
+ * carries `status.*` and `lifecycle_state.*` properties. Construct one per
+ * workspace.
  */
 export class WorkspaceAgentTelemetry {
 	private observed: ObservedAgentState | undefined;
@@ -124,10 +133,10 @@ export class WorkspaceAgentTelemetry {
 			{
 				workspaceName: this.workspaceName,
 				agentName: agent.name,
-				fromStatus: previous?.status ?? INITIAL_STATE,
-				toStatus: agent.status,
-				fromLifecycleState: previous?.lifecycleState ?? INITIAL_STATE,
-				toLifecycleState: agent.lifecycle_state,
+				"status.from": previous?.status ?? INITIAL_STATE,
+				"status.to": agent.status,
+				"lifecycle_state.from": previous?.lifecycleState ?? INITIAL_STATE,
+				"lifecycle_state.to": agent.lifecycle_state,
 			},
 			previous ? { observedDurationMs: now - previous.observedAtMs } : {},
 		);

src/logging/httpRequestsTelemetry.ts

@@ -175,15 +175,15 @@ export class HttpRequestsTelemetry implements Disposable {
 					{ method, route },
 					{
 						window_seconds: elapsedSeconds,
-						count_1xx: bucket.count1xx,
-						count_2xx: bucket.count2xx,
-						count_3xx: bucket.count3xx,
-						count_4xx: bucket.count4xx,
-						count_5xx: bucket.count5xx,
-						count_network_error: bucket.countNetworkError,
-						p50_duration_ms: percentile(sortedDurations, 0.5),
-						p95_duration_ms: percentile(sortedDurations, 0.95),
-						p99_duration_ms: percentile(sortedDurations, 0.99),
+						"count.1xx": bucket.count1xx,
+						"count.2xx": bucket.count2xx,
+						"count.3xx": bucket.count3xx,
+						"count.4xx": bucket.count4xx,
+						"count.5xx": bucket.count5xx,
+						"count.network_error": bucket.countNetworkError,
+						"duration.p50_ms": percentile(sortedDurations, 0.5),
+						"duration.p95_ms": percentile(sortedDurations, 0.95),
+						"duration.p99_ms": percentile(sortedDurations, 0.99),
 					},
 				);
 			}

test/unit/api/coderApi.test.ts

@@ -256,12 +256,12 @@ describe("CoderApi", () => {
 				{ method: "GET", route: "/api/v2/workspaces/{id}" },
 				expect.objectContaining({
 					window_seconds: 60,
-					count_1xx: 0,
-					count_2xx: 1,
-					count_3xx: 0,
-					count_4xx: 0,
-					count_5xx: 0,
-					count_network_error: 0,
+					"count.1xx": 0,
+					"count.2xx": 1,
+					"count.3xx": 0,
+					"count.4xx": 0,
+					"count.5xx": 0,
+					"count.network_error": 0,
 				}),
 			);
 		});

test/unit/instrumentation/workspace.test.ts

@@ -118,13 +118,23 @@ describe("WorkspaceStateTelemetry.observe", () => {
 	it("emits the first observation with from=none and no duration", () => {
 		const { sink, instance: state } = setup(newState);
 
-		state.observe(createWorkspace({ latest_build: { status: "running" } }));
+		state.observe(
+			createWorkspace({
+				latest_build: {
+					status: "running",
+					transition: "start",
+					reason: "initiator",
+				},
+			}),
+		);
 
 		const event = sink.expectOne("workspace.state_transitioned");
 		expect(event.properties).toMatchObject({
 			workspaceName: WORKSPACE_NAME,
 			from: "none",
 			to: "running",
+			"build.transition": "start",
+			"build.reason": "initiator",
 		});
 		expect(event.measurements.observedDurationMs).toBeUndefined();
 	});
@@ -168,10 +178,10 @@ describe("WorkspaceAgentTelemetry.observe", () => {
 
 		expect(sink.expectOne("workspace.agent.state_transitioned")).toMatchObject({
 			properties: {
-				fromStatus: "none",
-				toStatus: "connecting",
-				fromLifecycleState: "none",
-				toLifecycleState: "created",
+				"status.from": "none",
+				"status.to": "connecting",
+				"lifecycle_state.from": "none",
+				"lifecycle_state.to": "created",
 			},
 		});
 	});
@@ -197,7 +207,7 @@ describe("WorkspaceAgentTelemetry.observe", () => {
 
 		const events = sink.eventsNamed("workspace.agent.state_transitioned");
 		expect(events).toHaveLength(2);
-		expect(events[1].properties.fromStatus).toBe("none");
+		expect(events[1].properties["status.from"]).toBe("none");
 	});
 
 	it("includes observedDurationMs between transitions", () => {

test/unit/logging/httpRequestsTelemetry.test.ts

@@ -80,22 +80,22 @@ describe("HttpRequestsTelemetry", () => {
 			{ method: "GET", route: "/api/v2/workspaces/{id}" },
 			{
 				window_seconds: WINDOW_SECONDS,
-				count_1xx: 0,
-				count_2xx: 2,
-				count_3xx: 0,
-				count_4xx: 0,
-				count_5xx: 0,
-				count_network_error: 0,
-				p50_duration_ms: 100,
-				p95_duration_ms: 200,
-				p99_duration_ms: 200,
+				"count.1xx": 0,
+				"count.2xx": 2,
+				"count.3xx": 0,
+				"count.4xx": 0,
+				"count.5xx": 0,
+				"count.network_error": 0,
+				"duration.p50_ms": 100,
+				"duration.p95_ms": 200,
+				"duration.p99_ms": 200,
 			},
 		);
 		expect(log).toHaveBeenNthCalledWith(
 			2,
 			"http.requests",
 			{ method: "POST", route: "/api/v2/workspaces/{id}" },
-			expect.objectContaining({ count_2xx: 1 }),
+			expect.objectContaining({ "count.2xx": 1 }),
 		);
 	});
 
@@ -115,12 +115,12 @@ describe("HttpRequestsTelemetry", () => {
 			"http.requests",
 			{ method: "POST", route: "/api/v2/users/{name}/workspaces" },
 			expect.objectContaining({
-				count_1xx: 1,
-				count_2xx: 1,
-				count_3xx: 1,
-				count_4xx: 1,
-				count_5xx: 1,
-				count_network_error: 1,
+				"count.1xx": 1,
+				"count.2xx": 1,
+				"count.3xx": 1,
+				"count.4xx": 1,
+				"count.5xx": 1,
+				"count.network_error": 1,
 			}),
 		);
 	});
@@ -160,9 +160,9 @@ describe("HttpRequestsTelemetry", () => {
 			"http.requests",
 			{ method: "GET", route: "/api/v2/workspaces/{id}" },
 			expect.objectContaining({
-				p50_duration_ms: 100,
-				p95_duration_ms: 190,
-				p99_duration_ms: 200,
+				"duration.p50_ms": 100,
+				"duration.p95_ms": 190,
+				"duration.p99_ms": 200,
 			}),
 		);
 	});
@@ -187,10 +187,10 @@ describe("HttpRequestsTelemetry", () => {
 			"http.requests",
 			{ method: "GET", route: "/api/v2/workspaces/{id}" },
 			expect.objectContaining({
-				count_2xx: 1,
-				p50_duration_ms: 0,
-				p95_duration_ms: 0,
-				p99_duration_ms: 0,
+				"count.2xx": 1,
+				"duration.p50_ms": 0,
+				"duration.p95_ms": 0,
+				"duration.p99_ms": 0,
 			}),
 		);
 	});
@@ -208,7 +208,7 @@ describe("HttpRequestsTelemetry", () => {
 		expect(log).toHaveBeenCalledWith(
 			"http.requests",
 			{ method: "GET", route: "/api/v2/workspaces/{id}" },
-			expect.objectContaining({ count_2xx: 1 }),
+			expect.objectContaining({ "count.2xx": 1 }),
 		);
 	});
 

test/unit/remote/workspaceStateMachine.test.ts

@@ -445,16 +445,16 @@ describe("WorkspaceStateMachine", () => {
 			const events = sink.eventsNamed("workspace.agent.state_transitioned");
 			expect(events).toHaveLength(2);
 			expect(events[0].properties).toMatchObject({
-				fromStatus: "none",
-				toStatus: "connecting",
-				fromLifecycleState: "none",
-				toLifecycleState: "created",
+				"status.from": "none",
+				"status.to": "connecting",
+				"lifecycle_state.from": "none",
+				"lifecycle_state.to": "created",
 			});
 			expect(events[1].properties).toMatchObject({
-				fromStatus: "connecting",
-				toStatus: "connected",
-				fromLifecycleState: "created",
-				toLifecycleState: "ready",
+				"status.from": "connecting",
+				"status.to": "connected",
+				"lifecycle_state.from": "created",
+				"lifecycle_state.to": "ready",
 			});
 			expect(events[1].measurements.observedDurationMs).toEqual(
 				expect.any(Number),
@@ -489,16 +489,16 @@ describe("WorkspaceStateMachine", () => {
 			const events = sink.eventsNamed("workspace.agent.state_transitioned");
 			expect(events).toHaveLength(2);
 			expect(events[0].properties).toMatchObject({
-				fromStatus: "none",
-				toStatus: "connected",
-				fromLifecycleState: "none",
-				toLifecycleState: "ready",
+				"status.from": "none",
+				"status.to": "connected",
+				"lifecycle_state.from": "none",
+				"lifecycle_state.to": "ready",
 			});
 			expect(events[1].properties).toMatchObject({
-				fromStatus: "none",
-				toStatus: "connecting",
-				fromLifecycleState: "none",
-				toLifecycleState: "created",
+				"status.from": "none",
+				"status.to": "connecting",
+				"lifecycle_state.from": "none",
+				"lifecycle_state.to": "created",
 			});
 		});
 	});

test/unit/workspace/workspaceMonitor.test.ts

@@ -109,14 +109,14 @@ describe("WorkspaceMonitor", () => {
 				properties: {
 					from: "running",
 					to: "stopping",
-					transition: "stop",
-					reason: "autostop",
+					"build.transition": "stop",
+					"build.reason": "autostop",
 				},
 				measurements: { observedDurationMs: expect.any(Number) },
 			});
 		});
 
-		it("dedupes on (status, transition, reason); re-emits when only reason changes", async () => {
+		it("dedupes on (status, build transition, build reason); re-emits when only reason changes", async () => {
 			const { stream, sink } = buildSinkContext();
 
 			await setup(
@@ -153,7 +153,7 @@ describe("WorkspaceMonitor", () => {
 
 			const reasons = sink
 				.eventsNamed("workspace.state_transitioned")
-				.map((e) => e.properties.reason);
+				.map((e) => e.properties["build.reason"]);
 			expect(reasons).toEqual(["autostop", "initiator"]);
 		});