ark trace
Look up a specific task trace in a running Jaeger or Tempo instance.
ark trace <task-id> [flags]
Flags
| Flag | Default | Description |
|---|---|---|
--endpoint | $JAEGER_ENDPOINT | Base URL of the Jaeger or Tempo UI |
--format | tree | Output format: tree or json |
Examples
# Jaeger
ark trace 1710000001-0 --endpoint http://localhost:16686
# Tempo
ark trace 1710000001-0 --endpoint http://tempo.monitoring.svc.cluster.local:3100
# Use environment variable
export JAEGER_ENDPOINT=http://localhost:16686
ark trace 1710000001-0
# JSON output
ark trace 1710000001-0 --format json
Finding task IDs
Task IDs are stored in ArkTeam.status.steps[].taskID:
kubectl get arkteam research-pipeline -n my-org \
-o jsonpath='{.status.steps[*].taskID}'
# 1710000001-0 1710000045-0 1710000089-0
Example output
Task 1710000001-0 — research-pipeline/researcher
Duration: 4.2s
Status: Succeeded
Tokens: in=320 out=1240
Trace:
arkonis.task [4.2s]
├─ arkonis.llm.call [3.8s] model=llama3.2 in=320 out=1240
└─ arkonis.tool.call [0.1s] tool=web-search__search
Local traces with ark run --trace
When running locally, --trace collects spans in-memory and prints the tree without any backend:
ark run quickstart.yaml --trace
Flow succeeded in 9.1s — 1,516 tokens
arkonis.task [9.1s]
├─ arkonis.llm.call [4.2s] research in=1,204 out=312
└─ arkonis.llm.call [2.1s] summarize in=312 out=88
See also
- Observability concept — OTel traces setup and Jaeger deployment
- Environment Variables reference —
OTEL_EXPORTER_OTLP_ENDPOINT