The purpose of this document is to provide enhanced visibility into the status of edge clusters during Day-0 and Day-N operations. This granularity aims to empower users with better information and facilitate easier debugging through the Orchestrator.
Configurational States
Once an edge cluster, comprising one or more nodes, is created, it undergoes a boot-up process for initialization. The steps involved in this process are displayed in the Edge Clusters table, under the Status column.
Booting Up
Initializing the node boot-up process
Getting Serial Number
Verifying the presence of a serial number.
Enabling Access
Activating teleport access for remote connectivity.
Configuring Data Interface
Checking for an assigned IP address on the data interface.
Operations Booting Up
Communicating with devops to gather information, such as node role (leader vs. follower).
Connecting to Orchestrator
Establishing communication with the Orchestrator to register the node and acquire certificates for subsequent API calls.
Setting Up Control Plane
Initializing Kubernetes on the node.
Configuring Secure Network
Establishing networking services within Kubernetes.
Configuring Cluster Monitoring
Installing the metrics server for collecting node CPU/Memory usage metrics.
Setting Up Datastore
Installing MongoDB on the edge.
Services Initializing
Installing the remaining Celona services.
Node Operational
Indicates successful deployment when all Kubernetes pods are up, and node boot-up is complete.
Note: Steps 8, 9, and 10 are only applicable for leader nodes.
Operational States
Once the boot-up is complete, the cluster enters an operational state, displaying either an Up or Down status. The criteria for a cluster to be marked as Up are as follows:
Leader node in the cluster must be Up
Majority of the nodes in the cluster must be Up.
If these conditions are not met, the cluster is marked as Down
Understanding the Operational Status for Single-Node Clusters
For single-node clusters, the operational status is determined by the following criteria:
If the Leader node is down, the cluster is marked as Down.
If the Leader node is Up, the cluster is marked as Up.
Node-level status during boot-up is replicated at the cluster level for single-node clusters, ensuring alignment between individual node status and overall cluster status.
Understanding the Operational Status for Three-Node Clusters
For three-node clusters, the operational status is determined by the following criteria:
If the Leader node is Up, and the majority of the nodes are Up, the cluster is marked as Up.
If the majority of the nodes are not up (2/3 are in the initialization state), the cluster is marked as Down.
If the majority of the nodes are down, the cluster is marked as Down.
If the Leader node is Up, and the majority of the nodes are Up, the cluster is considered Up. However, if one or more follower nodes are down, the cluster operates at reduced capacity, denoted by a yellow warning tooltip. This indicates that the cluster is Up but operating with reduced capacity due to one or more node outages.
Note: Please ensure the Orchestrator is on version 2308 or higher to view these cluster-level states.