When Firebase Goes Down: A Survival Guide for Your Team
Your app loads instantly — the Firebase Hosting layer serves the static shell in under 200ms. But users see a loading spinner where their data should be. They click buttons and nothing happens. Firestore is returning 503s in us-central1. The app looks alive but nothing works.
What Happens on Your Team
The Mobile Developer
Opens the app on their phone — it loads, shows the UI skeleton, but data never populates. Checks the Firebase console — Firestore reads are failing. Cloud Functions are timing out. But the Firebase status dashboard shows all green for their region.
The real cost: Firebase apps often fail partially — the shell loads but functionality breaks. Users see what looks like a buggy app, not a service outage. They blame your app quality, not the infrastructure.
What they should have had: A monitor on a Cloud Function health endpoint that reads from Firestore and returns 200. When Firestore is down, the function fails, the monitor catches it, and you know the backend is broken before users assume your app is buggy.
The Frontend Developer
Deploys a new version to Firebase Hosting. The deploy succeeds, hosting serves the new files. But the new code calls a Cloud Function that hasn't been deployed yet. Users get the new UI with broken functionality.
The real cost: Firebase Hosting and Cloud Functions deploy independently. A mismatch between frontend and backend versions breaks the app silently — there's no build error, no deployment failure, just a runtime mismatch that only monitoring catches.
What they should have had: A post-deploy monitor check that verifies the full stack works — not just that the page loads, but that API calls succeed. Set up a monitor hitting your Cloud Function's endpoint within minutes of deployment.
The Startup CTO
Gets a user report: "Your app was broken all weekend." Checks Firebase — everything looks fine now. Checks usage metrics — there's a clear dip in API calls from Friday evening to Monday morning. Realizes a Firestore regional outage affected their project for 48 hours with no alert.
The real cost: Firebase doesn't proactively notify you about project-level issues. Platform incidents appear on the status page, but regional degradation, quota exhaustion, and cold start spikes are invisible without external monitoring.
What they should have had: External monitoring with weekend alerting. A Telegram alert on Saturday morning would have triggered an investigation and a status page update for users — instead of discovering the outage on Monday.
Why Monitor Firebase?
Firebase services are independent — Firestore can be down while Auth works fine, or Cloud Functions can timeout while Hosting serves pages normally. Since Firebase apps typically depend on multiple services simultaneously, a partial outage breaks your app in ways that are hard to diagnose without monitoring each piece.
What to Monitor
your-app.web.appYour Firebase Hosting URLyour-app.firebaseapp.comDefault Firebase app domainus-central1-your-project.cloudfunctions.net/healthCloud Functions health endpointWhat You Should Actually Do
- 1Monitor your Firebase Hosting URL to catch deployment failures and CDN issues
- 2Create a Cloud Function health endpoint that tests Firestore reads — this catches backend outages that don't affect the static shell
- 3Track response times on Cloud Functions — cold start spikes are common and can make your app feel broken during low-traffic periods
- 4Set up alerts for weekends and off-hours — Firebase outages don't respect your work schedule
- 5Monitor your custom domain separately from your .web.app domain — DNS and SSL issues on custom domains won't affect the default Firebase domain
Firebase's Official Status Page
Firebase publishes real-time status at status.firebase.google.com. Monitoristic doesn't replace this — it complements it. The official page tells you when Firebase reports an issue. Your own monitor tells you when your connection is affected, often before the status page updates. You also get push alerts instead of checking a webpage manually.
Firebase apps have a unique failure mode: the static shell loads perfectly while the backend is broken. Users see your app, interact with your UI, and think it's buggy — not that the infrastructure is down. External monitoring that tests actual functionality, not just page loading, is the only way to catch these partial failures.
Related Reading
Skip the panic. Know in 60 seconds.
Start Monitoring Firebase →Plans from $5/month · 14-day money-back guarantee