In the context of Rack Resiliency, the term “critical services” refers to those Kubernetes services which RR is monitoring and trying to make resilient. RR supports Kubernetes Deployments and StatefulSets as critical services.
HPE has prepopulated the RR critical services with the set of services necessary for the successful execution of user jobs. Note that this does not include all services that could reasonably be considered critical to the system overall. For example, the Boot Orchestration Service (BOS) and the Power Control Service (PCS) are both necessary for booting compute nodes. However, because they are not necessary for executing user jobs, they are not included in the default RR critical services list. It possible to add additional critical services to the list. For more information on managing the critical services, see Manage Critical Service.
Internally, the RR API service and the Resiliency Monitoring Service manage the critical services using a set of Kubernetes ConfigMaps. For more details, see ConfigMaps.