Rescue from network errors when querying Prometheus in the background
What does this MR do and why?
Currently PrometheusAdapter
can fail to update its reactive cache when encountering unhandled network errors.
This is because only a subset of all possible errors in Gitlab::HTTP::HTTP_ERRORS
is handled by Gitlab::PrometheusClient#get
.
This MR rectifies this by making PrometheusClient
reraise all HTTP_ERRORS
as PrometheusClient::Error
.
Screenshots or screen recordings
Refer to: #341633 (comment 714590480)
How to set up and validate locally
Refer to: #341633 (comment 714590480)
MR acceptance checklist
This checklist encourages us to confirm any changes have been analyzed to reduce risks in quality, performance, reliability, security, and maintainability.
-
I have evaluated the MR acceptance checklist for this MR.
Related to #341633 (closed)
Edited by Dominic Bauer