Check agent health in detail

Full node health: datapath mode, IPAM, BPF maps, controllers, the first thing to run when something breaks

cilium status --verbose

tool
cilium23 commands
tags
ciliummonitoringtroubleshootingk8s

More cilium commands

  • hubble observe --protocol tcp --from-pod <ns>/<pod> --to-fqdn '*.example.com'

    Live L3/L7 flows from a pod to a DNS name, useful to check FQDN-based egress policy

  • cilium clustermesh status --wait

    Check connectivity between clusters and the sync of remote endpoints and identities

  • cilium-dbg bpf nat list | head

    List BPF masquerade and NAT entries to debug SNAT port exhaustion or the egress source IP

  • cilium status --wait

    Check that all Cilium components are healthy and wait until they are ready

  • cilium connectivity test

    Run built-in end-to-end connectivity tests

all 23 cilium commands