Should we move our database to RDS or keep it on self-managed EC2?

[GregoryZL]

I’m trying to decide if we should move our main application database to a managed service like Amazon RDS or keep it on our self-managed EC2 instances. The promised reduction in operational overhead is tempting, but I’m nervous about losing fine-grained control over performance tuning and backup schedules.

[AndrewFJ]

We moved to a managed service like RDS last year and the overhead dropped in practice—less patching, fewer surprise maintenance windows, automatic backups, and easier failover tests. The flip side was that some fine grained knobs I used to tweak on the fly disappeared or required workarounds. We learned to work with the SLA and the metrics the vendor exposes rather than chasing every single DB setting.

[OliviaXH]

I keep wondering if the real bottleneck isn’t the DB at all. We spent weeks tweaking connections, buffers, and cache sizing and still had latency during peak hours. Maybe profiling the app would have helped more than fiddling with DB parameters.

[Steven55]

Phased migration helped. We kept core writes on EC2 and moved reads to a managed service, then added a small cache layer. We could monitor in one place, and when we needed tighter control we adjusted through the vendor's dashboard.

[GregoryL]

Drift off topic for a moment, but I started a DR drill log and kept stumbling over snapshot timing and retention. Odd how something so boring ends up swallowing cycles. Still, that forced a re-think about backups and where the real risk sits.