Most outages are found by customers, not by the team that runs the site. This checklist covers the twelve things worth watching so that you hear about problems first.
Availability
1. Your home page and key pages
Check that your most important URLs return a successful response, not just that the server is reachable. A page that loads an error message can still answer with a healthy status, so also check for text you expect to see.
2. Critical journeys and APIs
Login, checkout, sign-up and your public API are where failures cost the most. Monitor those endpoints with the same request a real client would send, and define the response that counts as healthy.
3. Response time
A site that answers in eight seconds is technically up and practically broken. Record a baseline, then alert when response time stays above the level your visitors would tolerate.
4. More than one location
A single vantage point cannot tell a regional network problem from a real outage. Checking from several locations helps you separate the two and reduces false alarms.
Expiry and configuration
5. SSL certificates
An expired certificate shows visitors a full-page security warning. Track expiry dates and ask for reminders well before the deadline, not on the day.
6. Domain registration
A lapsed domain takes everything offline, including email. Monitor expiry and make sure the renewal contact is someone who still works with you.
7. DNS records
An unexpected change to an A, MX or TXT record can redirect traffic or break email delivery. Watch your records and keep a history of changes so you can see what changed and when.
Infrastructure and jobs
8. Server resources
CPU, memory and disk rarely fail without warning. A disk that fills up slowly is one of the most avoidable causes of downtime, so alert on usage limits rather than waiting for errors.
9. Scheduled jobs and backups
Cronjobs fail silently. Have each job send a heartbeat when it finishes, and alert when an expected check-in does not arrive. Backups deserve this more than anything.
10. Page speed
Track performance scores over time so a slow release does not go unnoticed. Trends matter more than a single test result.
People and communication
11. Alert routing
Send alerts to the channels your team actually reads, and decide who is responsible for each one. Too many noisy alerts are as risky as too few.
12. A status page
When something does go wrong, a status page lets customers see what is happening without opening a support ticket.
Start small. If you do nothing else, monitor your home page, your SSL certificate and your domain expiry. Those three cover a surprising share of preventable incidents.
Watch your website the way this guide describes.
Start with the free Explorer plan and add checks as your needs grow.
Get Started with Bryzora ↗