#reliability
5 posts tagged reliability.
- SLOs for features that answer differently every timeLLM-backed features break the usual assumptions: slow by design, a dependency you do not run, and "success" that a status code cannot see. What to measure instead.
- Kubernetes requests are a purchase orderEvery CPU and memory request reserves capacity you pay for, whether the pod uses it or not. Get them right and the cluster gets smaller and more stable.
- Reserved concurrency is the best circuit breaker in AWSMost people meet Lambda concurrency as a limit to raise. Used deliberately, it is a dial that protects your database, isolates noisy functions and works as a kill switch.
- Certificates now last 200 days. Then 47.Public TLS certificates are on a schedule to shrink to 47 days by 2029. If renewal still involves a person, the schedule is the incident.
- The most-deployed untested code in your company is BashBash runs your entrypoints, your deploys and your backups, and almost nobody reviews it like code. Five failure modes that strict mode does not catch, and a line for when to stop.