From fd7720d5baf9ffecde42b4bb70f6165d25cb276b Mon Sep 17 00:00:00 2001
From: Gero Posmyk-Leinemann <gero@gitpod.io>
Date: Mon, 29 Aug 2022 16:02:23 +0000
Subject: [PATCH] [server] Only alert on explicit system failures

---
 components/server/src/prometheus-metrics.ts        |  6 +++++-
 .../server/src/workspace/workspace-starter.ts      | 14 +++++++-------
 2 files changed, 12 insertions(+), 8 deletions(-)

diff --git a/components/server/src/prometheus-metrics.ts b/components/server/src/prometheus-metrics.ts
index 1b3e79df359458..206d9787ad96e3 100644
--- a/components/server/src/prometheus-metrics.ts
+++ b/components/server/src/prometheus-metrics.ts
@@ -148,7 +148,11 @@ const instanceStartsFailedTotal = new prometheusClient.Counter({
     registers: [prometheusClient.register],
 });
 
-export type FailedInstanceStartReason = "clusterSelectionFailed" | "startOnClusterFailed" | "other";
+export type FailedInstanceStartReason =
+    | "clusterSelectionFailed"
+    | "startOnClusterFailed"
+    | "imageBuildFailed"
+    | "other";
 export function increaseFailedInstanceStartCounter(reason: FailedInstanceStartReason) {
     instanceStartsFailedTotal.inc({ reason });
 }
diff --git a/components/server/src/workspace/workspace-starter.ts b/components/server/src/workspace/workspace-starter.ts
index 108db903785f38..473944d2b4d040 100644
--- a/components/server/src/workspace/workspace-starter.ts
+++ b/components/server/src/workspace/workspace-starter.ts
@@ -591,12 +591,11 @@ export class WorkspaceStarter {
             if (rethrow) {
                 throw err;
             } else {
-                log.error("error starting instance", err, { instanceId: instance.id });
-                let failedReason: FailedInstanceStartReason = "other";
+                TraceContext.setError({ span }, err);
+                log.error({ userId: user.id, instanceId: instance.id }, "error starting instance", err);
                 if (err instanceof StartInstanceError) {
-                    failedReason = err.reason;
+                    increaseFailedInstanceStartCounter(err.reason);
                 }
-                increaseFailedInstanceStartCounter(failedReason);
             }
 
             return { instanceID: instance.id };
@@ -765,7 +764,7 @@ export class WorkspaceStarter {
         ideConfig: IDEConfig,
         pvcEnabledForPrebuilds: boolean,
     ): Promise<WorkspaceInstance> {
-        const span = TraceContext.startSpan("buildWorkspaceImage", ctx);
+        const span = TraceContext.startSpan("newInstance", ctx);
         //#endregion IDE resolution TODO(ak) move to IDE service
         // TODO: Compatible with ide-config not deployed, need revert after ide-config deployed
         delete ideConfig.ideOptions.options["code-latest"];
@@ -1264,7 +1263,7 @@ export class WorkspaceStarter {
 
             TraceContext.setError({ span }, err);
             const looksLikeUserError = (msg: string): boolean => {
-                return msg.startsWith("build failed:");
+                return msg.startsWith("build failed:") || msg.startsWith("headless task failed:");
             };
             if (looksLikeUserError(message)) {
                 log.debug(
@@ -1272,10 +1271,11 @@ export class WorkspaceStarter {
                     `workspace image build failed: ${message}`,
                 );
             } else {
-                log.warn(
+                log.error(
                     { instanceId: instance.id, userId: user.id, workspaceId: workspace.id },
                     `workspace image build failed: ${message}`,
                 );
+                err = new StartInstanceError("imageBuildFailed", err);
             }
             this.analytics.track({
                 userId: user.id,