Troubleshooting
Common setup and ingestion issues when connecting AutoObserve.
Docs v1.0Updated Supported
When should I use this?
- Telemetry or investigations are not behaving as expected after install or connect.
- You need a starting checklist before escalating to a design-partner session.
- You want quick links to the relevant OpenTelemetry or deploy guides.
Requirements
- AutoObserve deployed or design-partner environment provisionedRequired
- OpenTelemetry exporter configured (or in progress)Optional
Steps
1. Identify your symptom
Use the diagnostic assistant below to jump to the matching flow, or scan the symptom table.
| Symptom | Check |
|---|---|
| No investigations appear | Telemetry reaching AutoObserve? Time range correct? |
| Missing Kubernetes context | Cluster permissions and event/change sources |
| Incomplete RCA | Are metrics, logs, traces, and deploys all connected? |
| High resource use | Collector sampling and AutoObserve retention settings |
| Metrics/logs arrive, traces missing | Trace pipeline exporters include AutoObserve? |
| Partial correlation / weak identity | service.name, service.version, and Kubernetes attributes aligned? |
| Collector connected, no signals | Pipeline service section maps receivers → processors → exporters |
2. Follow the matching resolution flow
Each flow ends with a verification step so you know the fix worked.
Did it work?
Troubleshooting
Problem
No investigations appear after setup.
Symptoms
- Empty investigation list for known incident
Likely causes
- No telemetry ingested
- Incorrect time range
Resolution
Start with [Connect OpenTelemetry](/docs/connect/opentelemetry) and confirm verification.
Verify fix
Investigation appears for the target service and window.
Problem
Investigations lack Kubernetes context.
Symptoms
- No deploy or pod restart evidence
Likely causes
- Cluster permissions
- Missing k8s attributes on telemetry
Resolution
See [Kubernetes deployment](/docs/deploy/kubernetes) and resource attribute alignment.
Verify fix
Kubernetes evidence visible in investigations.
Problem
Root-cause analysis is incomplete.
Symptoms
- Hypothesis lacks supporting evidence types
Likely causes
- Missing metrics, logs, traces, or deploy signals
Resolution
Connect all signal types. See [Investigation model](/docs/investigate/overview).
Verify fix
Causal chain includes inspectable evidence across signal types.
Problem
Metrics and logs arrive but traces are missing.
Symptoms
- Trace pipeline empty in verification
Likely causes
- Trace exporters omit AutoObserve
Resolution
Add AutoObserve to the trace pipeline. See [Verify telemetry](/docs/connect/opentelemetry#verify-telemetry).
Verify fix
Traces receiving in verification checklist.
Problem
Correlation confidence is low across signals.
Symptoms
- Same service under different names
Likely causes
- Misaligned resource.attributes
Resolution
Align identity fields. See [Resource attributes](/docs/connect/opentelemetry#resource-attributes).
Verify fix
Consistent service identity in Observability.
Problem
Collector connected but no signals arrive.
Symptoms
- All verification channels empty
Likely causes
- Pipeline mapping error
- Auth or endpoint misconfiguration
Resolution
Validate pipeline service section and credentials in [OpenTelemetry docs](/docs/connect/opentelemetry).
Verify fix
At least one signal type receiving.
What's next?
Continue
15-minute Quick Start
Re-run the full deploy → connect → verify → investigate path from the beginning.
For design-partner deployments, escalate directly through your engineering working session rather than a public ticket queue.