CockroachDB High Memory, CPU or OOM Crash Fix

How to fix it

// try these in order, the most common fix is first

1. FIX 1Size memory flags

   Copy steps
   1. 01Start nodes with --cache=.25 --max-sql-memory=.25 (recommended for production)
   2. 02In Kubernetes, set limits and requests so the values match
   Did this fix it? Yes No

2. FIX 2Fix expensive queries

   Copy steps
   1. 01Check SQL Activity for high memory statements
   2. 02Add indexes and LIMITs; avoid huge result sets
   Did this fix it? Yes No

3. FIX 3Limit connections

   1. 01Use a connection pool and cap max connections per node (guidance: about 4x vCPUs)
   Did this fix it? Yes No

4. FIX 4Scale

   1. 01Add nodes or vCPUs; schedule backups and schema changes off-peak
   Did this fix it? Yes No

Symptoms

Nodes killed by OOM; memory climbs; CPU pinned at 100%; 'memory budget exceeded' errors.

Why it happens

• C-01--cache and --max-sql-memory left at defaults on a shared or containerized host
• C-02Large sorts, joins and aggregations in memory
• C-03Too many open connections
• C-04Heavy background jobs (backups, schema changes, CDC)
• C-05Container memory limits lower than the settings

// this guide also covers

• cockroachdb out of memory
• cockroachdb high memory usage
• cockroachdb high cpu
• cockroachdb memory budget exceeded
• cockroachdb oom killed

Other CockroachDB errors

• CockroachDB 'restart transaction' Error (40001) Fix
• CockroachDB Cluster Throttled: License Expired or No Telemetry
• Why Is CockroachDB Slow? Slow Query Fixes
• CockroachDB Connection Refused / Cannot Connect Fix
• CockroachDB Node Down or Cluster Unavailable: Fix
• CockroachDB 'Not Supported' / Postgres Compatibility Errors Fix
• CockroachDB Cloud Bill Too High: How to Cut Costs
• How to Cancel or Downgrade CockroachDB Cloud