Production n8n stability
Stop crashes before migrating the database.
A diagnostic for high-stakes self-hosted n8n systems moving from MySQL to Postgres without losing workflows, credentials, execution history, or confidence.
First-pass investigation map
| Risk area | What I verify before touching production |
|---|---|
| Crash source | Container restarts, memory pressure, execution bloat, queue mode settings, DB locks, worker concurrency, and disk pressure. |
| Backup safety | Current MySQL dump, n8n data volume, encryption key, env file, workflows, credential metadata, and restore rehearsal path. |
| Migration path | Schema compatibility, execution-history retention choice, dry-run import, credential validation, and rollback window. |
| High-load architecture | Postgres, Redis, queue mode, workers, pruning policy, restart policy, health checks, and reverse proxy limits. |
| Critical workflow proof | One government/API/financial routing workflow tested end-to-end after migration before broad traffic returns. |
$29 diagnostic deliverable
- Redacted evidence checklist: logs, Docker Compose, env names, DB size, execution count, and crash timeline.
- Root-cause hypothesis ranked by risk and proof needed.
- Zero-data-loss migration plan from MySQL to Postgres with rollback points.
- Stability hardening checklist for queue mode, workers, pruning, backups, and monitoring.
- Fixed-scope estimate for the migration and stabilization execution.