BREAKING: Supabase MINOR 🚨 pgBouncer issues on older projects [6ky4qpn3srzg]

Supabase experiencing pgBouncer connection pooling failures on legacy projects. Immediate workarounds and monitoring steps for indie hackers.

BREAKING: Supabase pgBouncer Issues Affecting Older Projects

Status: Investigating | Severity: MINOR | Impact: Partial disruption

What's Down & Who's Affected

Supabase is reporting pgBouncer connection pooling failures on select older projects—primarily those created before Q3 2023. The issue manifests as:

Connection timeouts when querying PostgreSQL

FATAL: remaining connection slots reserved for non-replication superuser connections errors

Intermittent 502/503 responses on API calls using pooled connections

Database operations hanging or failing after 30+ seconds

Who's impacted: Indie projects on legacy infrastructure, older starter/pro tier deployments. Newer projects are unaffected. Not all old projects are hit—this is partial.

Immediate Workarounds (Do This NOW)

1. Switch to Direct Connections

Temporarily bypass pgBouncer: ```

Instead of your pooled connection string:

postgres://user:pass@aws-0-region.pooler.supabase.com/postgres

Use direct connection:

postgres://user:pass@aws-0-region.supabase.co/postgres ``` Note the .pooler.supabase.com → .supabase.co change. You'll lose connection pooling benefits but restore immediate access.

2. Reduce Connection Pool Size

If you must use pooler, lower your pool size: ```javascript // Reduce from default 10-15 to 3-5 const pool = new Pool({ max: 3, // was 10 idleTimeoutMillis: 30000 }); ```

3. Implement Retry Logic

Add exponential backoff in your client: ```javascript const retryQuery = async (query, maxRetries = 3) => { for (let i = 0; i < maxRetries; i++) { try { return await db.query(query); } catch (err) { if (i === maxRetries - 1) throw err; await new Promise(r => setTimeout(r, 1000 * Math.pow(2, i))); } } }; ```

4. Scale Read Replicas

Route read traffic to read replicas if you have them configured.

Check If Your Project Is Affected

1. Go to Supabase Dashboard → Your project 2. Settings → Database → Copy your connection string 3. Look for .pooler.supabase.com in the URL 4. Test direct connection by replacing .pooler with .supabase.co 5. If that works but pooler doesn't = you're affected

Alternatively, test via psql: ```bash psql "postgres://user:pass@aws-0-region.pooler.supabase.com/postgres" -c "SELECT 1"

Compare with:

psql "postgres://user:pass@aws-0-region.supabase.co/postgres" -c "SELECT 1" ```

Alternative Tools to Consider

If this becomes chronic:

Neon: Serverless Postgres with built-in connection pooling (no pgBouncer issues)

Railway: Simple managed Postgres, newer infrastructure

Render: Good for smaller projects, reliable pooling

Fly Postgres: If you're already in the Fly ecosystem

*Don't migrate yet—this should be resolved in hours.*

Monitor Recovery

1. Check official status: https://status.supabase.com (refresh every 5 min) 2. Test your pooler connection every 15 minutes 3. Watch Supabase Discord: #incidents channel for updates 4. Set alerts: Use your monitoring tool to ping pooler endpoint 5. Follow up: Supabase will post RCA (Root Cause Analysis) post-incident

Bottom Line

Use direct connections NOW. This is temporary. The team is actively investigating. Older projects should see resolution within 2-4 hours. New projects are unaffected. Keep your logs on for debugging.

Stay calm. This is a connection pool issue, not data loss.