Insights:How to Configure Domain & DNS Tools: From Alerts to On-Call
Layer 1: DNS Resolution Monitoring
At a minimum, monitor three types of records:
- Whether the A/AAAA records for the root domain and
wwwreturn the expected IPs. - Whether NS records still point to the correct providers.
- Whether MX and critical TXT records are still present.
Monitoring frequency doesn't need to be very high — once every 5 to 15 minutes is sufficient. The key is multi-location probing: ping from only one data center, and you won't catch regional resolution issues.
Layer 2: Certificate Monitoring
Certificates are the easiest to overlook because they expire at midnight. What to monitor:
- Reminders at 30 days, 14 days, and 7 days before expiration.
- Whether the certificate chain is complete.
- Whether both the root domain and
wwware covered. - Whether HTTPS handshakes are succeeding normally.
Don't send expiry alerts by email alone; also push them to the on-call chat group.
Layer 3: Availability Monitoring
Normal DNS resolution doesn't guarantee the site is reachable. Add an external HTTP check:
- The homepage should return 200, not 302 redirect to an error page.
- Response times have a baseline — alert when they degrade suddenly.
- Key pages (login, forms, product pages) checked separately.
Don't use a single monitor for all pages. Configure critical paths individually, sample regular pages.
Layer 4: Alert Routing
An alert that no one watches is as good as no alert at all. At minimum, define:
- Primary recipient: who handles it during business hours.
- Backup recipient: who takes over if the primary doesn't respond.
- Escalation rules: after how long without action, escalate to the manager.
- Silence window: how to avoid false positives during planned maintenance.
- Alert content: include domain, record type, current resolved value, and last change.
Alert copy should drive action—don't just write "DNS abnormal."
Layer 5: Change Records
Each time you modify NS, A, CNAME, MX, TXT records, or certificates, log an entry:
- Who made it and when.
- The value before and after the change.
- Why the change was needed.
- How to verify and how to roll back.
Without change records, you're left guessing during incidents. A shared spreadsheet is plenty; no need for a complex system.
On-Call Scheduling
Small teams don't need full 7×24 coverage. You can structure it as:
- Business hours: fixed responsible person.
- Off-hours: rotating on-call with one primary and one backup.
- Holidays: check certificates and expiry dates in advance; avoid migrations over breaks.
The on-call schedule should clearly state which alert level routes to whom—don't send everything to the boss.
Run a Drill Before Going Live
At least simulate once:
- Manually tweak a test record and see how quickly monitoring detects it.
- Check who receives the alert and whether the content supports troubleshooting.
- Simulate a certificate expiration reminder and confirm someone handles it.
- Roll back following the change record and time the process.
After one drill cycle, real incidents won't start with "who changed this?" in the group chat.
Tool Selection
Pick tools that support multi-location probing, HTTPS and certificate checks, and deliver alerts to actual people. Free tools cover basic needs; there's no rush to buy an enterprise plan upfront. The key is having someone own it and respond.
Related Reading: