Bug reports
Send Runlot a problem with the evidence attached — the failing requests, worker exits, issues and your last failed command — instead of a screenshot of a 502.
runlot bug-reportWith no flags it asks for a one-line description and attaches the linked project's failing requests, worker exits and open issues from the last hour. Before anything is sent it prints what will be attached — counts and a preview — and asks you to confirm.
runlot bug-report --request 4bf92f3577b34da6a3ce929d0e0e4736 --message "checkout returns 502 since v42"
runlot bug-report --since 30m
runlot bug-report --last| Flag | What it does |
|---|---|
--request <id> | Attach a request by its ID — from the Runlot-Request-Id header, an error page or the dashboard. Repeat it, up to 20 |
--since <window> | The window for failing requests, worker exits and issues: 30m, 2h, up to 24h. Default 1h |
--message "…" | The description. Asked for in a terminal when you leave it out; required when there is no terminal |
--last | Attach the most recent runlot command on this machine that failed |
--no-logs | Attach request records and worker exits only, no log lines |
--dry-run | Print exactly what would be sent and stop |
--yes, -y | Send without asking |
When a runlot command fails with a server error or something unexpected, it ends with a hint to run runlot bug-report --last.
What is sent
The CLI sends references, not log contents: request IDs, a time window, and with --last the failed command. Runlot resolves them and keeps a snapshot with the report, so the evidence outlives the 3- or 7-day retention:
- the requests you named, with their logs, spans and the process exit that explains them;
- failing requests and worker exits in the window;
- open issues in Runlot Errors;
- the current and previous deployment, and recent builds.
From your machine it adds the CLI version, OS, architecture and Node version, and your runlot.json. Values that look like secrets, or sit under secret-like names, are replaced — once by the CLI and again by the server. With --last, the command's arguments are reduced to the command words, with every value removed, plus its exit status, the end of its error output and the request IDs of the API answers it got.
After you send it
You get an email with the report's ID (BR-…); quote it in any follow-up. Replies arrive by email and on the report. Any member can file a report about a project they can see. Organization admins can list the organization's reports and delete one, which deletes its snapshot.
Each time Runlot staff read the snapshot, it appears in the project's audit trail.