Issue

Metrics, logs, or traces are not being collected or sent to the OpsRamp portal, even though the Kubernetes 2.0 agent pods are running.

Cause

1. Pipelines not Started
   • The OpenTelemetry (OTel) pipelines for metrics, logs, or traces did not start properly.
2. Configuration Errors
   • Missing or incorrect configuration in receivers, processors, exporters, connectors, or extensions may prevent signals from being collected or exported.
3. Corrupted or Incomplete Pipeline Files
   • The OTel configuration files under /opt/opsramp/agent/tmp/otel/infra may be missing or incorrectly set up.

Solution

1. Check Agent Logs
   • Review the agent.log file to confirm if the metrics, logs, and traces pipelines have started successfully.
   • Look for sections such as:
     • Receivers in the config
     • Exporters in the config
     • Processors in the config
     • Connectors in the config
     • Extensions in the config
     • Pipelines in the config
     • Telemetry in the config
2. Exec into Agent Pod
   • Access any worker or master agent pod:

     kubectl exec -it <opsramp-worker-pod-name> -- sh

3. Inspect Pipeline Files
   • Navigate to the pipeline configuration directory:

     cd /opt/opsramp/agent/tmp/otel/infra

   • List the available configuration files:

     ls -la

4. Verify File Contents
   • Use cat to open and review configuration files:

     cat podlogs-receiver.yaml
     cat logs-opsrampotlp-exporter.yaml

   • Ensure each receiver, processor, and exporter has correct settings (for example, exporter endpoints must point to the OpsRamp portal).
5. Restart Agent if Needed
   • If pipeline configuration was corrected, restart the agent pods:

     cat podlogs-receiver.yaml
     kubectl rollout restart deployment <opsramp-agent-deployment> -n <namespace>

6. Validate in Portal
   • Check the OpsRamp portal to confirm that metrics, logs, and traces are flowing.