When Heroku apps go down, the first question is always: is this a Heroku platform outage or an application error? Heroku's error code system (H10, H12, H14, R14...) provides clues, but distinguishing platform issues from app bugs requires a systematic approach. This guide walks you through the diagnosis.
How to Check if Heroku is Down (Step by Step)
1. Check the Official Heroku Status Page
Visit status.heroku.com. Heroku breaks down status by component:
- Runtime: Dyno execution and routing
- API: Heroku Platform API and CLI operations
- Dashboard: Web management console
- Deploys: Git push and GitHub integration deploys
- Add-ons: Heroku Postgres, Redis, Kafka, and third-party add-ons
2. Use the Heroku CLI for Quick Diagnostics
The Heroku CLI provides several diagnostic commands:
# Check platform status heroku status # View recent app logs heroku logs --tail -a your-app-name # Check dyno status heroku ps -a your-app-name # Check Postgres status heroku pg:info -a your-app-name
3. Read the Error Code
Heroku's error codes tell you a lot before you even check the status page:
| Error Code | Meaning | Likely Cause |
|---|---|---|
| H10 | App Crashed | Application startup error — check logs |
| H12 | Request Timeout | App didn't respond within 30 seconds |
| H13 | Connection Closed Without Response | App closed socket before response |
| H14 | No Web Dynos Running | Dyno formation scaled to 0 |
| H80 | Maintenance Mode | Platform maintenance or incident |
| R14 | Memory Quota Exceeded | App using too much RAM |
| R15 | Memory Quota Vastly Exceeded | Dyno killed for excessive memory use |
📡 Monitor Heroku uptime every 30 seconds — get alerted in under a minute
Trusted by 100,000+ websites · Free tier available
Why Does Heroku Go Down? Common Causes
- AWS Infrastructure Issues: Heroku runs on AWS. AWS us-east-1 is Heroku's primary region — AWS incidents there often affect Heroku operations.
- Routing Mesh Degradation: Heroku's HTTP router is a high-traffic system. Degradation can cause H80 errors or slow response times across all apps in a region.
- Heroku Postgres Incidents: Shared database plans can experience resource contention. Dedicated and premium plans have better isolation.
- Deploy Pipeline Failures: GitHub integration or Heroku CI failures during Heroku platform updates can block deployments even when apps run normally.
- Add-on Provider Outages: Third-party add-ons (Papertrail, Redis Cloud, Sendgrid) have independent availability. An add-on outage isn't a Heroku platform failure.
Monitor your Heroku apps with external uptime checks
Better Stack monitors your Heroku app URLs from multiple global regions. Get instant alerts when response time spikes or apps go offline — regardless of where the failure originates.
Try Better Stack Free →Diagnosing App Crashes vs. Platform Outages
The most common mistake: assuming Heroku is down when the issue is application-level. Use this diagnostic flow:
- Run
heroku status— if it shows issues, it's platform-level. Stop here. - Check
heroku ps -a [app]— are dynos in a crashed state? Look for "crashed" or restarting dynos. - Run
heroku logs --tail -a [app]— look for startup errors in the last 100 lines. - Test a canary request — use
curl -I https://your-app.herokuapp.com. What status code do you get? - Check resource usage —
heroku pg:infofor database connections;heroku redis:infofor Redis.
Action Plan: What to Do When Heroku is Down
For Active Incidents:
- Don't attempt redeployments during a platform incident — they may worsen state.
- Communicate to your users via your status page (if you have one).
- Check if Heroku's incident affects your specific region (US, EU) or is global.
For Production Apps with SLAs:
- Consider migrating critical production workloads to Railway, Render, or direct AWS/GCP if Heroku outages are a recurring problem.
- Use Heroku Private Spaces for better SLAs and isolation.
- Maintain a blue-green deployment setup to quickly switch traffic during incidents.
Store your Heroku API keys and environment variables securely
Heroku config vars contain API keys, database URLs, and secrets. 1Password CLI integrates with Heroku to inject secrets without storing them in .env files or git history.
Try 1Password Free →Heroku Alternatives Worth Evaluating
If Heroku downtime is affecting your business, consider evaluating platforms with better availability track records:
- Railway: Modern Heroku alternative with more predictable pricing and strong uptime. Supports the same Git-push deploy workflow.
- Render: Popular Heroku migration destination with free tiers, auto-deploy from GitHub, and managed Postgres.
- Fly.io: Deploy Docker containers globally with built-in health checks and auto-restarts.
- AWS Elastic Beanstalk: Closer to infrastructure control with familiar Heroku-style deployment. Better for teams ready to manage more config.
Frequently Asked Questions
My Heroku app shows H10 errors — is Heroku down?
H10 (App Crashed) usually means your application crashed on startup, not a platform failure. Check heroku logs --tail -a [app] for the actual error. Common causes: missing environment variables, dependency installation failures, syntax errors in startup scripts. Only if many apps are affected and status.heroku.com shows an incident is it a platform issue.
Heroku deployments are failing — what should I check?
First check status.heroku.com for any "Deploys" component incidents. Then check: (1) your git push output for specific errors, (2) heroku builds:info for build log details, (3) your Procfile and runtime.txt for configuration issues, (4) buildpack compatibility with your new dependencies.
How do I get notified when Heroku has an outage?
Subscribe to email or webhook notifications at status.heroku.com. Follow @HerokuStatus on X for real-time incident updates. For automated monitoring of your specific apps, configure external uptime monitoring via Better Stack or API Status Check.
Alert Pro
14-day free trialStop checking — get alerted instantly
Next time Heroku goes down, you'll know in under 60 seconds — not when your users start complaining.
- Email alerts for Heroku + 9 more APIs
- $0 due today for trial
- Cancel anytime — $9/mo after trial