Better error handling UX during pipeline creation
Problem to solve
Although we surface expected errors during the pipeline creation process in the form of a failed pipeline, for unexpected errors we never surface that an error occurred and instead show an endless spinner. For context: Unexpected errors could be network related.
Proposal
Record somewhere(redis, database) pipeline creation failures for a given ref and surface an error if the pipeline could not be created due to an unexpected error instead of a endlessly spinning. Since the creation process auto-retries up to 3 times we could surface the error on the first unexpected error and then not that it is retrying until it no longer is? @veethika, do you have design ideas around this?
Edited by Allison Browne