This is a useful reality check. Most of production support is not dramatic outages; it is the same small set of problems coming back: noisy alarms, a cost spike nobody can explain, a deploy that only one person knows how to run.
One approach that has worked for our clients is treating each recurring issue as its own small, fixed project rather than an endless support retainer. For example, alarms tuned so every page is actionable, or a cost dashboard with tags that answer who spent what. Each one has a clear definition of done, and after a few of them the support load drops noticeably.
Which category takes most of your on call time today?
This is a useful reality check. Most of production support is not dramatic outages; it is the same small set of problems coming back: noisy alarms, a cost spike nobody can explain, a deploy that only one person knows how to run.
One approach that has worked for our clients is treating each recurring issue as its own small, fixed project rather than an endless support retainer. For example, alarms tuned so every page is actionable, or a cost dashboard with tags that answer who spent what. Each one has a clear definition of done, and after a few of them the support load drops noticeably.
Which category takes most of your on call time today?