Executive Summary
Overall: INCONCLUSIVE
The local OTEL emission loop and Grafana are both passing. The current run (26047160379) is confirmed visible in both the local JSONL mirror and Grafana Tempo. Sentry's write path also shows zero export errors, and the gh-aw project exists in the github Sentry org — but the available Sentry MCP tools only expose error-event queries, not OTEL span/trace queries, so current-run span visibility in Sentry cannot be confirmed via this toolset.
Status Table
| Backend |
Write Config Present |
Write Export Succeeded |
Read Config Present |
Read Query Succeeded |
Overall |
| Local OTLP |
✅ |
✅ |
— |
— |
✅ |
| Sentry |
✅ |
✅ |
✅ |
🔶 |
🔶 |
| Grafana |
✅ |
✅ |
✅ |
✅ |
✅ |
Evidence
Local OTLP
- OTEL env vars:
OTEL_EXPORTER_OTLP_ENDPOINT=set, OTEL_EXPORTER_OTLP_HEADERS=set, GH_AW_OTLP_ENDPOINTS=set, OTEL_SERVICE_NAME=gh-aw.smoke-otel-backends
/tmp/gh-aw/otel.jsonl: 1 line present
- Current-run span:
{"name":"gh-aw.agent.setup","run_id":"26047160379"} confirmed
- OTLP export errors:
0
Sentry
- MCP connection: ✅ Connected to
https://us.sentry.io (org: github)
- Project found: ✅
gh-aw project present in Sentry org github
- Current-run spans: 🔶 Cannot confirm —
list_events only surfaces error events, not OTEL spans/traces. No errors in project (expected for a healthy workflow).
- Recent gh-aw spans: 🔶 Not queryable via available MCP tools (no
search_spans or performance trace query tool exposed)
- OTLP write errors attributable to Sentry: none (export error count = 0)
Grafana
- MCP connection: ✅ Connected to Grafana Cloud (
grafanacloud-ghaw-* stack)
- Tracing datasource: ✅
grafanacloud-traces (UID: grafanacloud-traces, type: tempo)
- Current-run spans: ✅ Trace
a0b383979b3661ce56581a69f042e0dd found with github.run_id=26047160379, root span gh-aw.pre_activation.setup
- Recent gh-aw spans: ✅ Additional traces found (e.g.
6fe3741ea748d3b85241747b96f5f730, root gh-aw.activation.setup)
- Representative trace:
a0b383979b3661ce56581a69f042e0dd — 26 311 ms, service gh-aw.smoke-otel-backends
Failure Analysis
Sentry — Read Query Inconclusive
Failing step: Read-side query for OTEL spans in Sentry.
Evidence observed: The Sentry MCP server exposes 12 tools. None of them provide a span/trace/performance query surface (search_spans, get_trace, or equivalent are absent). list_events queries the errors dataset and returned zero results — correct for a workflow that emits no errors. The gh-aw project exists and the write path shows 0 OTLP export errors, so data may be ingested but is not reachable via this read path.
Most likely root cause: The Sentry MCP tool configuration (sentry.json) does not include a performance/tracing query tool. The Sentry project may be configured only for error ingestion, or the OTEL spans land in Sentry's performance/tracing product which is not surfaced in this MCP integration.
Path classification: Read path — tooling/configuration gap, not a backend failure.
Next concrete debug step:
- Check whether Sentry's
gh-aw project has "Performance" / "Tracing" enabled under Project Settings → Performance.
- Verify that the Sentry OTLP endpoint in
GH_AW_OTLP_ENDPOINTS points to the traces ingest URL (e.g. (o/redacted)*.ingest.sentry.io/api/*/envelope/ or the new OTLP endpoint).
- Add a
search_spans or get_trace tool to the Sentry MCP configuration so future smoke runs can confirm read-side visibility.
Run URL: https://github.com/github/gh-aw/actions/runs/26047160379
Generated by 🧪 Smoke OTEL for issue #33044 · ● 4.2M · ◷
Executive Summary
Overall: INCONCLUSIVE
The local OTEL emission loop and Grafana are both passing. The current run (
26047160379) is confirmed visible in both the local JSONL mirror and Grafana Tempo. Sentry's write path also shows zero export errors, and thegh-awproject exists in thegithubSentry org — but the available Sentry MCP tools only expose error-event queries, not OTEL span/trace queries, so current-run span visibility in Sentry cannot be confirmed via this toolset.Status Table
Evidence
Local OTLP
OTEL_EXPORTER_OTLP_ENDPOINT=set,OTEL_EXPORTER_OTLP_HEADERS=set,GH_AW_OTLP_ENDPOINTS=set,OTEL_SERVICE_NAME=gh-aw.smoke-otel-backends/tmp/gh-aw/otel.jsonl: 1 line present{"name":"gh-aw.agent.setup","run_id":"26047160379"}confirmed0Sentry
https://us.sentry.io(org:github)gh-awproject present in Sentry orggithublist_eventsonly surfaces error events, not OTEL spans/traces. No errors in project (expected for a healthy workflow).search_spansor performance trace query tool exposed)Grafana
grafanacloud-ghaw-*stack)grafanacloud-traces(UID:grafanacloud-traces, type:tempo)a0b383979b3661ce56581a69f042e0ddfound withgithub.run_id=26047160379, root spangh-aw.pre_activation.setup6fe3741ea748d3b85241747b96f5f730, rootgh-aw.activation.setup)a0b383979b3661ce56581a69f042e0dd— 26 311 ms, servicegh-aw.smoke-otel-backendsFailure Analysis
Sentry — Read Query Inconclusive
Failing step: Read-side query for OTEL spans in Sentry.
Evidence observed: The Sentry MCP server exposes 12 tools. None of them provide a span/trace/performance query surface (
search_spans,get_trace, or equivalent are absent).list_eventsqueries theerrorsdataset and returned zero results — correct for a workflow that emits no errors. Thegh-awproject exists and the write path shows 0 OTLP export errors, so data may be ingested but is not reachable via this read path.Most likely root cause: The Sentry MCP tool configuration (
sentry.json) does not include a performance/tracing query tool. The Sentry project may be configured only for error ingestion, or the OTEL spans land in Sentry's performance/tracing product which is not surfaced in this MCP integration.Path classification: Read path — tooling/configuration gap, not a backend failure.
Next concrete debug step:
gh-awproject has "Performance" / "Tracing" enabled under Project Settings → Performance.GH_AW_OTLP_ENDPOINTSpoints to the traces ingest URL (e.g.(o/redacted)*.ingest.sentry.io/api/*/envelope/or the new OTLP endpoint).search_spansorget_tracetool to the Sentry MCP configuration so future smoke runs can confirm read-side visibility.Run URL: https://github.com/github/gh-aw/actions/runs/26047160379