honeybadger-io/honeybadger

Creates, tunes, and explains Honeybadger Insights alarms, which watch a BadgerQL query and notify when its result count crosses a threshold. Use when the user wants to be alerted, paged, or notified about something in production ("let us know when 500s spike", "alert if the nightly job stops reporting"), asks to create or change an alarm, says an alarm is noisy, broken, or never fires, or asks what alarms exist.

Use when asked to investigate, debug, triage, or fix an error reported to Honeybadger — a fault link or ID, an error class or message, "the most recent error", or "what's broken in production".

Sets up Honeybadger in a project end to end, or orients a user whose project already uses it. Use whenever the user wants to start using Honeybadger, add Honeybadger or error tracking/monitoring to an app, install a Honeybadger client library, check whether their app reports to Honeybadger, or asks what Honeybadger can do from the agent — even if they don't say "get started".

Investigates a production problem with Honeybadger Insights events, faults, and deploys, and reports what happened, why, and who was affected. Use when the user brings a customer support ticket or complaint ("a customer says checkout failed"), says the app was slow or down, asks what happened during an incident or outage or for a postmortem, or asks why a metric changed. For a single known error (a fault link, ID, or error class) use honeybadger-debug-errors instead.

Adds a Honeybadger check-in to a scheduled job so the user is alerted when it stops running. Use when asked to monitor a cron job, scheduled task, rake task, periodic script, or background worker with Honeybadger, to add a check-in or ping to a job, or to find out when a job "goes silent" or "stops running". Also use right after adding a new scheduled job to a project that already uses Honeybadger, to offer a check-in for it.