Analytics

ElasticClaw Server records persistent analytics for workflows so you can see which automations are running, how often they succeed, and what happens after agents open pull requests.

Open Settings → Analytics in the web UI to review workflow performance.

• Total triggers — how many times workflows fired
• Success rate — percentage of agent creations that succeeded
• PRs opened — number of pull requests created by workflow agents
• PR merge rate — percentage of opened pull requests that merged
• Run statuses — breakdown of clean, human_in_the_loop, warning, failed, and running runs
• Cost and usage — estimated AI spend and token usage per model, run, and time range
• Run history — per-workflow run records with status, trigger type, timestamps, and linked agent
• Recent events — the latest workflow events with timestamps

Runs are classified into one of five statuses based on their outcome and whether a human intervened.

• clean — the run completed successfully and the PR merged, or no PR was required
• human_in_the_loop — the run completed but a human comment or action influenced the outcome
• warning — the run completed but a post-completion event (e.g. PR closed unmerged) was recorded
• failed — the run stopped because of an error, timeout, or required gate failure
• running — the run is still active

Analytics data is retained for up to 1 year. Use the time range selector to view the last 7, 30, or 90 days.