Improve fault tolerance of CI jobs

[shonfeder]

Its fairly common for jobs to fail with flaky errors (e.g. because of network issues). To clean up the reliability of the CI results, we want to start making these operations that can have flaky faults tolerant by adding retries and such.

Please add any flaky errors you encounter to this issue.

---

• git-fetch failed with exit-code 128 : Improve fault tolerance of CI jobs #391 (comment)

[shonfeder]

From ocaml/opam-repository#26916 (comment)

2024-11-20 09:24.38: Got resource from pool OCluster
Building on m1-worker-03
error: RPC failed; curl 92 HTTP/2 stream 7 was not closed cleanly: CANCEL (err 8)
error: 10845 bytes of body are still expected
fetch-pack: unexpected disconnect while reading sideband packet
fatal: early EOF
fatal: fetch-pack: invalid index-pack output
git-fetch failed with exit-code 128