Beyond Status Codes
A 200 OK doesn't mean your API is healthy. We've seen plenty of endpoints return 200 while serving stale data, timing out for 80% of users, or silently dropping errors in logs.
Here are 5 things you should be monitoring:
1. Response Time
A slow API is often a broken API in disguise. Monitor:
- p50 latency: Your typical response time
- p95 latency: The experience of your slower 5% of users
- p99 latency: Rare but catastrophic slowdowns
2. SSL Certificate Expiry
SSL certs expire without warning. We recommend monitoring:
- Days until expiry
- Certificate chain validity
- Proper HTTPS negotiation
3. DNS Resolution Time
DNS is often the forgotten bottleneck:
- DNS lookup time
- DNSSEC validation status
- Multiple DNS provider redundancy
4. Response Body Content
Don't just check status codes—validate the response:
- Does the JSON have expected fields?
- Is the data fresh (timestamp fields)?
- Are error rates within normal ranges?
5. TCP Connection Time
Before HTTP even happens, TCP matters:
- Connection establishment time
- TLS handshake time
- Time to first byte
Hawks monitors all of these out of the box. Set up your first monitor at hawks.example.com and see what you've been missing.