DNSMentor

Runbooks

A client is offline and you think it is DNS

15 min to do · 3 related

What do I do when a client is down and DNS might be the cause?

Establish whether DNS is actually the problem before changing anything: query the authoritative nameservers directly, compare with a public resolver, and check the last applied change in the audit trail. Most DNS-shaped outages are either a change made minutes ago or a delegation that expired, and both are visible in under two minutes.

Do not start changing records

The most common way a ten-minute DNS incident becomes a day-long one is somebody changing records to see whether it helps, while the previous answers are still cached. Diagnose first.

# The first two minutes

  1. Check the audit trail Open the domain in DNSMentor and look at recent changes. If something was applied in the last few hours, that is your first suspect and you have the exact diff.
  2. Ask the authoritative nameservers directly If the authoritative answer is correct, the problem is caching or something downstream — not the zone.
Authoritative answer, then what the world sees
dig @ns1.provider.net contoso.co.uk A +short
dig @1.1.1.1 contoso.co.uk A +short
dig contoso.co.uk NS +trace
  1. Check the delegation at the parent A +trace shows whether the parent zone still points at the nameservers you think it does. Expired domains and registrar changes both present as a total outage.
  2. Check whether it is mail specifically If web works and mail does not, look at MX first — and remember mail queues rather than bounces, so "mail is down" often means a change made up to five days ago.

# If a recent change caused it

  • Roll it back from the change record rather than hand-editing. The inverse is composed from what actually changed, and it goes through the checks.
  • Accept that the rollback is subject to the TTL of the bad record. If it was published with a long TTL, resolvers will hold it.
  • Tell the client the expected recovery window in terms of that TTL rather than guessing.
Raise it as urgent

A ticket raised from inside the platform with "client currently offline" is treated as urgent on every plan, including trials. Include the domain and the change reference.

Was this helpful?

Related

Still stuck?

Raise a ticket from inside the platform and your tenant, plan and recent activity come attached automatically. A client currently offline is treated as urgent on every plan.

Last updated