Files generated per run

Every sync or download writes the following files into the output directory alongside the data files.

File	When written	Contents
hivel_agent.log	Every run	Full structured log of the run. Useful for debugging failures.
diagnostics.json	Every run	NDJSON - agent version, run args, per-step timing, memory and disk usage at completion.
<provider>/status.json	Every run	Sync window start/end, success/failure status, error message if any. Written per provider (e.g. jira/status.json, github/gh-main/status.json).
<provider>/summary.json	Every run	Entity counts per provider (e.g. issues fetched, PRs fetched).
debug.log	--debug-requests only	Raw HTTP request/response traffic to the source API. Use to diagnose API auth or connectivity issues.

Where is the output directory for `sync`?

Unlike download and work-metrics, sync has no --output-dir flag at all. Run it as-is (the normal cron case) and the agent creates a fresh temp directory per run instead (e.g. $TMPDIR/hivel-agent-<timestamp>/) - its path isn't logged anywhere, so it's not practical to find afterward. hivel_agent.log's contents are also mirrored to stdout, so redirect stdout/stderr in your cron/systemd entry if you want a stable place to check what happened - see Checking agent status & logs.

Sending config to Hivel

Enable send_agent_config: true in config.yml to upload a sanitized copy of your config to Hivel on every run:

hivel:
  api_token: ${HIVEL_API_TOKEN}
  send_agent_config: true

All credential fields are redacted before upload (token, password, api_token → ***). Only the structural config (URLs, project filters, field mappings) is sent. This helps the Hivel support team diagnose configuration issues without you needing to share the file manually.

Snapshot Review

Configuration Reference