Troubleshooting VPS and Cloud Server Selection Issues: A 5-Step Guide

When a server fails, the worst approach is rebooting first, then reinstalling. The correct order is to confirm symptoms, check CPU/memory, disk, network and DNS…

Insights:Troubleshooting VPS and Cloud Server Selection Issues: A 5-Step Guide

Step 1: Confirm the Issue

  • Is the entire site down, or is a specific page slow?
  • All regions, or only certain ones?
  • Continuous outage, or intermittent?
  • Any error codes, logs, or monitoring alerts?
  • Any recent code changes, config updates, or certificate renewals?

Document first; do not touch production yet.

Step 2: System Resources

top
free -m
uptime

Check:

  • Whether CPU has been consistently maxed out.
  • Whether memory is fully consumed by cache or processes.
  • Whether system load exceeds the number of cores.
  • Whether any abnormal processes are running.

Step 3: Disk

df -h
df -i
iostat -x 1

Check:

  • Whether disk usage has exceeded 80%.
  • Whether inodes are exhausted.
  • Whether IO wait remains persistently high.
  • Whether logs have filled up the disk.

A full disk can cause both database and web service failures.

Step 4: Network & DNS

curl -I https://example.com/
dig +short A example.com
ss -lntp

Check:

  • Whether DNS resolution is correct.
  • Whether required ports are listening.
  • Whether firewall or security groups allow traffic.
  • Whether origin pull and CDN are functioning normally.
  • Whether SSL certificates have expired.

Step 5: Application & Database

  • Web server error logs.
  • PHP/runtime logs.
  • Database slow queries and connection count.
  • Whether the cache service is available.
  • Whether scheduled tasks are piling up.
  • Don't rely solely on system monitoring — application-level errors often hide in logs.

    Troubleshooting Record

    Time Symptom Check Result Action
    17:15 502 Disk 95% Clear logs / scale up
    17:20 502 Database Connections full Check slow queries

    Things Not to Do

    • Don't reinstall the OS as a first step.
    • Don't delete logs to cover up the issue.
    • Don't change production configs directly without keeping a backup.
    • Don't focus only on CPU and ignore disk and database.
    • Don't restart all services without a rollback plan.

    Locate the cause first, then fix it. Back up before acting.

    Related reading: