Confirm the outage

  1. Check the main domain and the exact failing page from an external checker.
  2. Try a second network or ask someone in another location.
  3. Record the error message, HTTP status and time.
  4. Check whether only logged-in users, checkout, email or another subsystem is affected.

A single failed browser is not enough evidence for a server outage. A failed health check from several networks is much stronger evidence.

Check what changed

Look at deployments, plugin or theme updates, DNS edits, certificate renewals, firewall rules and billing notices. Write down the last known working time and every change made after it.

If a deployment caused the outage and you have a tested rollback, one controlled rollback is usually safer than several speculative fixes.

Work through the service layers

Give support useful evidence

Send the provider the affected address, first known time with timezone, error text, HTTP status, steps already taken and whether the problem reproduces on another network. Avoid sending passwords, secret keys or customer data in an ordinary support message.

Communicate before certainty

A short factual notice is better than silence: name the affected function, say when you first noticed it, and promise the next update time. Do not invent a recovery estimate before the team responsible has one.

After recovery

Record the cause, the recovery action and one prevention step. Useful improvements may include external monitoring, tested backups, certificate reminders, a rollback procedure and a status page independent of the main host.