Closed ychen-atlassian closed 6 years ago
Also proposal for the same PR: How about expanding the error enums?
Forgot to answer. We need to rework how we propagate errors within Smith. We have https://github.com/atlassian/smith/issues/276 for this work, lets do it in another PR.
This adds a bunch of metrics into prometheus endpoint about bundle transitions, which would be useful for alerting purposes.
Also proposal for the same PR: How about expanding the error enums?
E.g. for resources we often just spit out the error but it would be useful to categorize if the Spec was wrong or service catalog didn't respond or if the resource couldn't be created.