Monitor and optimize a VPS

Measure CPU, memory, disk space, inodes and application response times before resizing the server or changing system settings.

Steps

1. Establish normal load and check basic capacity with read-only commands.

   uptime
   free -h
   df -h
   df -i
   top

2. Correlate slow requests with CPU saturation, swapping, disk latency, full filesystems and database delays. Linux filesystem cache is not automatically a memory leak.

3. Check service failures and out-of-memory events. Fix repeated restarts, unbounded queues and log growth.

4. Set alerts for remaining disk, sustained memory pressure, application errors and backup failures. Keep logs rotated without deleting active data blindly.

5. Profile and optimize the application before raising worker counts. Each worker and database connection consumes memory.

6. Review graph trends after a change. Schedule capacity upgrades when measured demand exceeds the safe operating range.

Check the result

A faster median response is not enough if errors or slowest-request times increase. Test recovery after a planned restart and verify a recent backup can actually be restored.

Related articles

• Performance guides
• Resources and disk
• Backups and recovery

Last reviewed

2026-10-28