Skip to main content

Analyze

drwn analyze sessions uploads a session-log archive to the configured analyzer backend and prints where to watch the job.

Preview without auth or network:

drwn analyze sessions --dry-run

Upload the newest local archive, or build one inline if none exists:

drwn analyze sessions

Build a fresh archive first:

drwn analyze sessions --fresh

Use an explicit upload artifact:

drwn analyze sessions --archive /tmp/sessions.tar.gz

Wait for the final report and open it:

drwn analyze sessions --wait --open

Emit machine-readable output:

drwn analyze sessions --wait --json

Configuration​

drwn analyze sessions requires auth from either drwn login or DRWN_TOKEN plus DRWN_ANALYZER_URL.

The analyzer API URL comes from credentials, DRWN_ANALYZER_URL, or analyzer.apiUrl. The optional frontend URL comes from DRWN_ANALYZER_WEB_URL or analyzer.webBaseUrl and is used to compose /processing/<jobId> and /report/<reportId> URLs.

{
  "version": 1,
  "analyzer": {
    "apiUrl": "http://localhost:8787",
    "webBaseUrl": "https://darwinian-harness-services.pages.dev",
    "maxArchiveBytes": 104857600
  },
  "optional": {}
}

Input resolution​

The command resolves input in this order:

  1. --archive <path> uses the explicit .tar, .tar.gz, or .tgz path.
  2. --fresh builds a new .tar.gz with the same discovery rules as drwn export sessions.
  3. The newest archive under .agents/drwn/session-log-exports/ is reused when present.
  4. If no archive exists, a new inline .tar.gz is built and uploaded.

--dry-run is non-mutating: it validates an existing archive when selected, or reports that an inline export would be built without creating it.

  • Export — build upload-ready archives manually
  • Login — authenticate before uploading