Group
  • Deployment

    Set up error alerting so you hear about bugs before users do

    Wire up an error tracker (Sentry or equivalent) with real grouping, notifications, and noise control.

    • Improve quality
    • Debug & fix
    • Beginner
    • General

Free Prompt

Set up an error tracking service for my app (Sentry, Rollbar, Bugsnag, Honeybadger, whatever I'm using or you recommend). Ask me if I already have one before installing anything. Wire it up properly: 1. Install the SDK on the server. Capture unhandled exceptions, unhandled promise rejections, and errors caught in top-level route handlers. Attach the request id, route, and user id (if authenticated) to every event so I can group by real dimensions. 2. Install the browser SDK too and capture client-side errors, but keep it lean (small bundle, sensible sampling). Same context: user id when signed in, current route, release version. 3. Set the release/version on each deploy so I can see which release introduced a new error. Enable source maps for the browser SDK so client stack traces are readable. 4. Configure notifications for real signals only: new issue in production, spike in error rate, regressions of resolved issues. Do not page me on every occurrence of a known issue; that's how alerts get muted and outages get missed. 5. Scrub PII: passwords, tokens, cookies, and email/user data I don't need for debugging should be filtered from breadcrumbs and request payloads. Log the user id, not the whole user object. Don't leave the default 'send everything' config in place; it will flood the free tier and turn every alert into noise. The deliverable is the SDK installed on server and client, release tracking wired to my deploys, PII filters configured, alert routing set to a channel I actually watch, and a test error captured end-to-end.

What This Does / How This Helps

This installs an error tracker with real context (request id, user, route, release), source maps for the browser, PII scrubbing, and alerts scoped to signals that matter instead of every single occurrence. Without error tracking, you find bugs the way your users do: by hearing about them in support tickets or angry tweets, hours after the fact. With it configured badly, you get so many notifications you stop reading them and miss the real one. Set up right, it's the first thing that tells you a bad deploy happened, usually before anyone has time to notice on their end.

Want to skip doing this by hand?

Fortivibe audits your app for all of the areas these prompts cover (and more).

See What We Check

Related Prompts