Health checks failing on EKS+overlay #26897
Description
Reproduction Steps
Start an EKS cluster. Then:
$ helm install cilium cilium/cilium --version 1.12.11 --namespace kube-system --set ipam.mode=cluster-pool --set tunnel=vxlan --set debug.enabled=true --set bpf.monitorAggregation=none
$ sleep 200 && ./cilium connectivity test --test-namespace c1 --test 'health'
[...]
[=] Test [health]
.⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
⚠️ cilium-health probe failed: "command terminated with exit code 1", stdout: {"" '\x00' '\x00'}, retrying...
I confirmed it affects Cilium v1.12.0--v1.12.11, v1.13.4, v1.11.18, and 1.14.0-rc.1.
This was known (cf. #16975 (comment)), but previous reports pointed to a closed issue that wasn't confirmed to be the same bug. Thus, nobody followed on it.