Our verifyInstrumentation job has been flickering lately, requiring a manual retry after which it succeeds.
There are a couple possible ways we could make this more efficient:
- Look into whether we can add automatic retries.
- Reduce 429 throttling and network timeouts from Maven by optimizing our verifier download strategy. Some ideas we discussed were (1) reducing the verifier ranges (similar to how we did for aws) (2) having some sort of way to track which versions we previously verified, and only verifying against new artifact versions (possibly with some sort of metadata file).
Our verifyInstrumentation job has been flickering lately, requiring a manual retry after which it succeeds.
There are a couple possible ways we could make this more efficient: