[hermes] metis #392: gitea_checkout_http_500 #352

New Issue

hermes-automation · 2026-09-30T13:28:23Z

hermes-automation commented

2026-09-30 13:28:23 +00:00

Hermes auto-triage classified incident metis/392 as gitea_checkout_http_500 (confidence 0.95); first failed gate: Declarative: Checkout SCM.

Why a human is needed

This build never actually got to your code: Jenkins failed while trying to git fetch the metis repository from the Gitea server (scm.bstein.dev), which returned an HTTP 500 instead of the expected pack data ('RPC failed; HTTP 500 ... fatal: expected packfile'). Jenkins retried the checkout per its normal retry policy and it kept failing, so it gave up before a workspace ever existed — that's also why no tests ran and no junit/metrics/artifacts were published, there was nothing to run against. There is no fix needed in the metis repository or its tests; this points to the Gitea/SCM host having an outage or overload at 2026-09-30T13:19–13:21 UTC. Re-running the same commit once the SCM host is responding normally should be enough to get a clean build.

Ariadne did not authorize automated remediation: classification_not_supported: 'gitea_checkout_http_500'.

Facts

jenkins — Jenkins reports the pipeline's first (and only) failed stage as 'Declarative: Checkout SCM'. (jenkins.first_failed_stage)
jenkins — The agent successfully provisioned and started (quality-tools container defined, JENKINS_AGENT_NAME assigned, remote channel connected); the failure occurred after agent startup during 'git fetch --tags --force --progress -- https://scm.bstein.dev/titan/metis.git', which returned status code 128 ... (jenkins.console_failures[1], line_number 215)
jenkins — Jenkins exhausted its checkout retry wrapper and aborted: 'ERROR: Error cloning remote repo 'origin'' / 'ERROR: Maximum checkout retry attempts reached, aborting'. (jenkins.console_tail)
jenkins — No structured test results were published (failed_tests is empty), and downstream post-actions (metrics publish, junit publish, artifact archive) were all skipped with 'workspace unavailable'. (jenkins.failed_tests and jenkins.console_tail)
opensearch — No corroborating log records were available; the OpenSearch query for this incident window failed with 'Connection refused'. (log_evidence.error)

Evidence

The earliest failure region of the build console:

tty: true
    volumeMounts:
    - mountPath: "/home/jenkins/agent"
      name: "workspace-volume"
  - command:
    - "cat"
    image: "registry.bstein.dev/bstein/quality-tools:sonar8.0.1-trivy0.70.0-db20260422-arm64"
    name: "quality-tools"
    tty: true
    volumeMounts:
    - mountPath: "/home/jenkins/agent"
      name: "workspace-volume"
  - env:
    - name: "JENKINS_SECRET"
      value: "********"
    - name: "JENKINS_TUNNEL"
      value: "jenkins.jenkins.svc.cluster.local:50000"
    - name: "JENKINS_AGENT_NAME"
      value: "metis-392-9gh5t-hn91t-5xfnw"

Inferences

The build never got past checking out the repository, so nothing in metis's own source or test suite ran or could have caused this failure.
The HTTP 500 came from the Gitea SCM host (scm.bstein.dev) while Jenkins was fetching refs, not from any client-side git configuration problem — curl 22 and 'fatal: expected packfile' both indicate the server sent back an error response instead of pack data.
Jenkins' built-in checkout retry already attempted this multiple times and still failed, implying the SCM host was erroring or unreachable for the whole retry window rather than hitting one single flaky request.

Links

Failed build: https://ci.bstein.dev/job/metis/392/
Full evidence bundle and audit trail live in Ariadne at /api/admin/audit/events, event types hermes_autotriage_incident and hermes_autotriage_diagnosis.

Filed automatically by Ariadne from a Hermes Agent diagnosis (run run_5700519fbe8648e3bcebc86df7fa2637). Hermes has no write access to this repository; no files or infrastructure were changed.

Hermes auto-triage classified incident `metis/392` as **gitea_checkout_http_500** (confidence 0.95); first failed gate: `Declarative: Checkout SCM`. ## Why a human is needed This build never actually got to your code: Jenkins failed while trying to git fetch the metis repository from the Gitea server (scm.bstein.dev), which returned an HTTP 500 instead of the expected pack data ('RPC failed; HTTP 500 ... fatal: expected packfile'). Jenkins retried the checkout per its normal retry policy and it kept failing, so it gave up before a workspace ever existed — that's also why no tests ran and no junit/metrics/artifacts were published, there was nothing to run against. There is no fix needed in the metis repository or its tests; this points to the Gitea/SCM host having an outage or overload at 2026-09-30T13:19–13:21 UTC. Re-running the same commit once the SCM host is responding normally should be enough to get a clean build. Ariadne did not authorize automated remediation: `classification_not_supported: 'gitea_checkout_http_500'`. ## Facts - **jenkins** — Jenkins reports the pipeline's first (and only) failed stage as 'Declarative: Checkout SCM'. (`jenkins.first_failed_stage`) - **jenkins** — The agent successfully provisioned and started (quality-tools container defined, JENKINS_AGENT_NAME assigned, remote channel connected); the failure occurred after agent startup during 'git fetch --tags --force --progress -- https://scm.bstein.dev/titan/metis.git', which returned status code 128 ... (`jenkins.console_failures[1], line_number 215`) - **jenkins** — Jenkins exhausted its checkout retry wrapper and aborted: 'ERROR: Error cloning remote repo \'origin\'' / 'ERROR: Maximum checkout retry attempts reached, aborting'. (`jenkins.console_tail`) - **jenkins** — No structured test results were published (failed_tests is empty), and downstream post-actions (metrics publish, junit publish, artifact archive) were all skipped with 'workspace unavailable'. (`jenkins.failed_tests and jenkins.console_tail`) - **opensearch** — No corroborating log records were available; the OpenSearch query for this incident window failed with 'Connection refused'. (`log_evidence.error`) ## Evidence The earliest failure region of the build console: ``` tty: true volumeMounts: - mountPath: "/home/jenkins/agent" name: "workspace-volume" - command: - "cat" image: "registry.bstein.dev/bstein/quality-tools:sonar8.0.1-trivy0.70.0-db20260422-arm64" name: "quality-tools" tty: true volumeMounts: - mountPath: "/home/jenkins/agent" name: "workspace-volume" - env: - name: "JENKINS_SECRET" value: "********" - name: "JENKINS_TUNNEL" value: "jenkins.jenkins.svc.cluster.local:50000" - name: "JENKINS_AGENT_NAME" value: "metis-392-9gh5t-hn91t-5xfnw" ``` ## Inferences - The build never got past checking out the repository, so nothing in metis's own source or test suite ran or could have caused this failure. - The HTTP 500 came from the Gitea SCM host (scm.bstein.dev) while Jenkins was fetching refs, not from any client-side git configuration problem — curl 22 and 'fatal: expected packfile' both indicate the server sent back an error response instead of pack data. - Jenkins' built-in checkout retry already attempted this multiple times and still failed, implying the SCM host was erroring or unreachable for the whole retry window rather than hitting one single flaky request. ## Links - Failed build: https://ci.bstein.dev/job/metis/392/ - Full evidence bundle and audit trail live in Ariadne at `/api/admin/audit/events`, event types `hermes_autotriage_incident` and `hermes_autotriage_diagnosis`. Filed automatically by Ariadne from a Hermes Agent diagnosis (run [run_5700519fbe8648e3bcebc86df7fa2637](https://triage.bstein.dev/chat?resume=run_5700519fbe8648e3bcebc86df7fa2637)). Hermes has no write access to this repository; no files or infrastructure were changed.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: titan/metis#352