flake: `allow-all/pod-to-service/curl-2` timeout during `curl` exec, no flows retrieved

[nbusseneau]

Hit in GKE on #390 https://github.com/cilium/cilium-cli/actions/runs/1001017290

2021-07-05T12:33:47.221179076Z   [.] Action [allow-all/pod-to-service/curl-2: cilium-test/client2-666976c95b-mc9mq (10.4.1.77) -> cilium-test/echo-other-node (echo-other-node:8080)]
2021-07-05T12:33:47.221185112Z   📄 Following flows...
2021-07-05T12:33:47.221856284Z   ❌ command "curl -w %***local_ip***:%***local_port*** -> %***remote_ip***:%***remote_port*** = %***response_code*** --silent --fail --show-error --connect-timeout 5 --output /dev/null http://echo-other-node:8080" failed: Timeout occurred
2021-07-05T12:33:47.221975647Z   📄 Validating flows for peer cilium-test/client2-666976c95b-mc9mq
2021-07-05T12:33:57.222499529Z   ❌ Aborting flow matching: context deadline exceeded
2021-07-05T12:33:57.222546717Z   ❌ Flow validation failed for peer cilium-test/client2-666976c95b-mc9mq: 0 failures (first: -1, last: -1, matched: 0)

Attaching sysdump: cilium-sysdump-out.zip.zip

List of PRs blocked:

• ~~#390~~ Manually unblocked after 2 weeks as an exception because it seems we have not reproduced this flake ever since.

[tklauser]

Hit by multicluster test in #529, though on no-policies/pod-to-service/curl-0, not allow-all/pod-to-service/curl-2: https://github.com/cilium/cilium-cli/runs/3547587561?check_suite_focus=true