System logs and metrics for Application Service Adapter

You can access Application Service Adapter system component logs and query metrics by following the steps in this topic.

Gather system logs

Logs for Application Service Adapter system components flow through standard Kubernetes logging channels. You can retrieve them by using the tooling of your choice. For ad-hoc log retrieval, VMware recommends the kapp logs command in the kapp cli as the most convenient method to gather the logs from pods in one or more deployments. See the Carvel documentation.

To fetch recent logs from all components of an Application Service Adapter installation:

kapp logs --app tas-adapter.app -n tap-install

To fetch recent logs from a specific deployment:

kapp logs --app tas-adapter.app -n tap-install --pod-name DEPLOYMENT-NAME%

Where DEPLOYMENT-NAME is the name of the Kubernetes Deployment, for example, korifi-api-deployment.

To stream logs instead of fetching the most recent logs, add the --follow flag to the earlier kapp logs commands. For additional details and options, see the kapp logs --help help text.

Query performance metrics

All of the controller-managers deployed by Application Service Adapter provide standard Prometheus performance metrics generated by the Kubernetes controller-runtime library. For more information, see the Kubebuilder documentation.

Application Service Adapter controller managers are deployed with annotations that makes them discoverable by a Prometheus server deployed to the same cluster.

After you have Prometheus deployed, you can query any of the available metrics. See Default Exported Metrics References in the Kubebuilder documentation. In particular, queries for workqueue_depth or workqueue_queue_duration_seconds can help to indicate when a controller manager is resource constrained and must be scaled up.

To query the queue depth for the CFApp controller:

workqueue_depth{name="cfapp"}

To list the queue depths for all Application Service Adapter controllers:

workqueue_depth{namespace="tas-adapter-system"}

To query the average queue time for CFApp objects before reconciliation:

workqueue_queue_duration_seconds_sum{name="cfapp"}/workqueue_queue_duration_seconds_count