NHC, SNR and FAR

2 views
Skip to first unread message

Nihar Ranjan Mishra

unread,
Jul 20, 2026, 11:13:34 AMJul 20
to medik8s

Or Raz

unread,
Jul 22, 2026, 3:25:58 AMJul 22
to Nihar Ranjan Mishra, medik8s

Hi Nihar,

Thanks for the question. The OpenShift installation page you linked lists ICMP generically as required for "Network reachability tests" between nodes, so it doesn't break down by specific ICMP message types.
Assuming we are speaking of the ICMP type messages 13 (Timestamp), 14 (Timestamp Reply), and 5 (Redirect), then NHC, SNR, or FAR are not using these kinds of ICMP messages.

Regarding the 3 Medik8s operators:

  1. Node Healthcheck Operator (NHC) - Does no network probing at all. It works entirely through the Kubernetes API, watching node.Status.Conditions and comparing them against configured thresholds. No ICMP of any kind.
  2. Fence Agents Remediation (FAR) - Delegates to external fence agent binaries (e.g., fence_ipmilanfence_aws). No ICMP of any kind.
  3. Self Node Remediation (SNR)- Operator uses three health-check mechanisms:
    1. API server check — HTTP GET to the API server's /readyz endpoint.
    2. Peer health check — gRPC over TLS between SNR agent pods.
    3. Endpoint health check (optional, control-plane only) — when endpointHealthCheckUrl is configured, SNR pings that endpoint using standard ICMP Echo (Type 8/0).

Is there a reason why you are asking about ICMP message types 13, 14, and 5 concerning Medik8s operators?

Best regards,

ORaz


--
You received this message because you are subscribed to the Google Groups "medik8s" group.
To unsubscribe from this group and stop receiving emails from it, send an email to medik8s+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/medik8s/8c464bb5-9725-4829-9770-1987b1cd5b30n%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages