Charmed Kubeflow charms Prometheus metrics

The following Charmed Kubeflow charms provide metrics that enable their monitoring. This page provides a high level overview of each charm’s metrics.

Contents:

• argo-controller
• dex-auth
• envoy
• istio-pilot
• jupyter-controller
• katib-controller
• kfp-api
• knative-eventing
• knative-serving
• metacontroller-operator
• minio
• seldon-controller-manager
• training-operator

Argo controller

See the upstream documentation on metrics provided by Argo controller.

To view them through the Prometheus UI at a certain point in time, query {juju_application="argo-controller"}.

Dex Auth

• a custom metric counting HTTP requests, described in its code
• Go runtime and process metrics for monitoring the controller itself
• gRPC server metrics

To view them through the Prometheus UI at a certain point in time, query {juju_application="dex-auth"}.

Envoy

Upstream documentation provides information on metrics (statistics) for the different parts of envoy:

• Listener
• HTTP connection manager
• Cluster manager
• Server
• Acces logs

To view them through the Prometheus UI at a certain point in time, query {juju_application="envoy"}.

Istio pilot

See the upstream documentation on metrics provided by Istio pilot.

To view them through the Prometheus UI at a certain point in time, query {juju_application="istio-pilot"}.

Jupyter controller

Jupyter controller provides:

• custom notebook-related metrics decsribed in its code
• Go runtime metrics for monitoring the controller itself
• controller runtime metrics

To view them through the Prometheus UI at a certain point in time, query {juju_application="jupyter-controller"}.

Katib controller

Katib controller provides:

• custom experiment-related metrics decsribed in its code
• Go runtime metrics for monitoring the controller itself
• controller runtime metrics

To view them through the Prometheus UI at a certain point in time, query {juju_application="katib-controller"}.

Kfp api

Kfp api provides:

• custom metrics related to its different sub-components, which are described in the following parts of its code:
  • resource-manager
  • experiment-server
  • job-server
  • pipeline-server
  • pipeline-upload
  • run-server
• Go runtime and process metrics for monitoring the controller itself

To view them through the Prometheus UI at a certain point in time, query {juju_application="kfp-api"}.

Knative eventing

See the upstream documentation on metrics provided by Knative eventing.

To view them through the Prometheus UI at a certain point in time, query {juju_application="knative-eventing"}.

Knative serving

See the upstream documentation on metrics provided by Knative serving.

To view them through the Prometheus UI at a certain point in time, query {juju_application="knative-serving"}.

Metacontroller operator

Metacontroller operator provides:

• custom metrics decsribed in its code
• Go runtime and process metrics for monitoring the controller itself
• controller runtime metrics

To view them through the Prometheus UI at a certain point in time, query {juju_application="metacontroller-operator"}.

Minio

See the upstream documentation on metrics provided by Minio.

To view them through the Prometheus UI at a certain point in time, query {juju_application="minio"}.

Seldon controller manager

See the upstream documentation on metrics provided by Seldon controller manager.

To view them through the Prometheus UI at a certain point in time, query {juju_application="seldon-controller-manager"}.

Training operator

Training operator provides:

• custom job-related metrics decsribed in its code
• Go runtime and process metrics for monitoring the controller itself
• controller runtime metrics

To view them through the Prometheus UI at a certain point in time, query {juju_application="training-operator"}.

Last updated 4 days ago.