CPU, memory and performance
Compare guest workload and platform metrics before changing resources.
Common symptoms
Slow server, high CPU, load average, out of memory, OOM, process killed, RAM, resize, latency
What to check
- Note when the slowdown occurs and which request or workload is affected. Check the server's CPU, memory and network charts for the same period and display timezone.
- Through SSH or Console, inspect guest processes, load, memory and relevant system logs. Check for OOM or killed processes, disk pressure and application limits before assuming more CPU will fix it.
- If resizing is appropriate, review the server's resize preview, available limits, price and required power changes. Save important data and check the resulting guest resources after the operation completes.
Common questions
A process was killed with OOM. What should I check?
Check guest memory use, the affected process and system logs. Reduce an unintended workload or adjust application limits first; review available memory resizing if the workload genuinely needs more resources.
Why do guest readings differ from the panel charts?
They may measure different intervals or resource views, and chart data can arrive later. Compare the same period and timezone, alongside guest process evidence, before drawing a conclusion.
Open the relevant page
Before contacting support
Include the server name, time window with timezone, workload, relevant chart and sanitized process or OOM messages. Remove passwords, tokens and personal data from logs.
Keep secrets out of support requests · Remove passwords, private keys, access tokens, TOTP codes, backup codes and full payment details from screenshots and logs. Share the server or order name, time with timezone and exact sanitized error.
This opens an editable ticket draft. Nothing is sent automatically.