5.4.7.3 OcnrfPodInCongestedState

Table 5-92 OcnrfPodInCongestedState

Field Details
Description 'The pod {{$labels.kubernetes_pod_name}} of service {{$labels.app_kubernetes_io_name}} is in Congested state'
Summary 'kubernetes_namespace: {{$labels.kubernetes_namespace}}, nrflevel:{{$labels.NrfLevel}}, podname: {{$labels.kubernetes_pod_name}}, timestamp: {{ with query "time()" }}{{ . | first | value | humanizeTimestamp }}{{ end }} : The pod is in Congested state'
Severity Major
Condition One or more pods of a service are in congested state. This could be due to CPU usage or Pending Message Count above configured thresholds. Currently this is applicable for NfSubscription service only.
OID 1.3.6.1.4.1.323.5.3.36.1.2.7082
Metric Used ocnrf_pod_congested_state
Recommended Actions The alert is cleared when the CPU usage or Pending Message Count goes below the configured thresholds for the congested state.

Note: The thresholds can be viewed using REST API.

Steps:

Reassess if the NRF is receiving additional traffic.

If this is unexpected, contact My Oracle Support.

  1. Refer to alert to determine which pod is receiving high traffic. It may due to a sudden spike in traffic. For example: When one mated site goes down, the NFs move to the given site. Check if NF is sending high number of updates, register, or deregister.
  2. Check the service pod logs on Kibana to determine the reason for the errors.
  3. If this is expected traffic, then the thresholds levels may need to be re-evaluated as per the call rate and reconfigured as mentioned in Oracle Communications Cloud Native Core, Network Repository Function REST Specification Guide.
Available in OCI No