The Move
Developer Emmanuel Asika published a custom PostgreSQL backup system to avoid Supabase's paid backup tiers. The implementation uses pg_dump, scheduled jobs, and object storage - standard tools that predate managed services.
The story matters because it's not isolated. Teams are increasingly self-hosting Supabase components (Edge Functions, databases) on platforms like Fly.io and Render. The pattern: start with managed services, hit a pricing inflection point, evaluate alternatives.
The Trade-Offs
Supabase's paid backups include point-in-time recovery, automated retention policies, and SLAs. Custom solutions trade these for cost savings and control. The calculus changes based on:
- Dataset size (where vendor pricing compounds)
- Recovery time objectives (DIY adds complexity)
- Compliance requirements (GDPR retention, audit trails)
- Team capacity (who maintains the scripts)
For small teams without dedicated ops, custom backups introduce risk. A missed cron job or misconfigured retention policy can mean data loss. For larger teams with PostgreSQL expertise, the cost-benefit math shifts.
The Tooling Context
PostgreSQL backup options have matured:
- pg_basebackup: Built-in, incremental-capable in Postgres 17
- pgBackRest: Full/incremental backups, WAL archiving, PITR support
- Barman: Enterprise-grade, compression, retention automation
All three are open source. The question isn't technical capability - it's operational overhead versus vendor margin.
What This Signals
This isn't an anti-Supabase story. It's a reminder that managed services optimize for simplicity, not cost at scale. The interesting pattern: developers are comfortable evaluating that trade-off explicitly.
Supabase raised $80M in Series B (2024) and competes in a ~$10B PostgreSQL market. Their backup pricing reflects standard SaaS economics. Teams hitting those tiers have options - self-hosting, alternative BaaS providers (Appwrite, PocketBase), or hybrid approaches.
The broader question for CTOs: where do you want vendor lock-in, and where do you want control? Backups sit at that intersection. They're critical infrastructure that's also commoditized technology.
Worth Noting
No contrarian takes emerged in coverage (HackerNoon, 4M+ monthly readers). That's telling. The developer community sees this as pragmatic cost optimization, not a referendum on Supabase's value proposition.
For enterprise teams: if you're running Supabase at scale, audit your backup costs against pgBackRest or Barman implementation effort. The math may surprise you. If you're not, the managed service is probably worth it.