Kubernetes | Agent Controller POD Health & Lifecycle

Overview

This article provides diagnostic procedures for verifying the operational state of Aembit Edge Components (Agent Controller and Agent Injector) within a Kubernetes cluster. Use this guide to identify if the components are running, healthy, and ready to manage workload traffic.

Relates To

• Deployment: Kubernetes (EKS, GKE, K8s Self-Hosted)

• Components: aembit-agent-controller, aembit-agent-injector

Cause (The Symptoms)

You should follow these steps if you observe any of the following:

• Pods are stuck in Pending or CrashLoopBackOff.

• kubectl get pods shows 0/1 READY.

• Error: Readiness probe failed: HTTP probe failed with statuscode: 401

Solution

Solution

1. Verify Pod Execution State List all pods in the aembit namespace to confirm they are in a Running state and have a 1/1 readiness ratio.

kubectl get pods -n aembit -o wide

ROSA / OpenShift:

oc get pods -n aembit -o wide

• Healthy Outcome: Both the Controller and Injector show STATUS: Running and READY: 1/1.

• Unhealthy Outcome: If pods show 0/1or a high restart count, proceed to step 2.

2. Inspect Lifecycle Events and Probes Retrieve the detailed events and configuration for the pods to identify why probes might be failing.

kubectl describe pods -n aembit

ROSA / OpenShift:

oc describe pods -n aembit

• Check: Scroll to the Events section. Look for warnings related to Liveness (Port 80) or Readiness (Port 443) probe timeouts.

3. Validate Configuration Environment Confirm that the Controller has been injected with the correct Aembit Tenant and Domain variables.

kubectl describe pods -n aembit | grep -A 10 "Environment"

ROSA / OpenShift:

oc describe pods -n aembit | grep -A 10 "Environment"

• Required Fields: Ensure AEMBIT_AGENT_CONTROLLER_ID and AEMBIT_TENANT_ID match your Aembit Management Portal settings.