Monitor AlloyDB Omni Kubernetes operator custom resources

This page describes how to monitor the status of AlloyDB Omni Kubernetes operator custom resources using kube-state-metrics (KSM).

Running KSM with the Kubernetes operator lets you do the following:

  • Monitor custom resource status in real time on your Prometheus-compatible observability stack, together with AlloyDB Omni metrics.
  • Build dashboards and alerts on top of the custom resource metrics for better operational visibility and proactive issue detection.

For more information about kube-state-metrics, see Overview. To learn how kube-state-metrics monitors Kubernetes custom resources, see Custom Resource State Metrics.

How AlloyDB Omni Kubernetes operator custom resources work

By default, custom resource metrics aren't enabled. You can enable custom resource metrics using the AlloyDB Omni Kubernetes operator Helm chart. When you enable custom resource metrics, a KSM deployment is created with a configmap that defines the custom resource metrics and a service of type ClusterIP to expose the metrics endpoint of the KSM deployment. Your Prometheus collector can collect the metrics, and you can visualize these metrics with other AlloyDB Omni metrics. For more information, see Custom resources metrics for the AlloyDB Omni Kubernetes operator.

Before you begin

Make sure that you have the following:

  • A Prometheus-compatible metrics scraper running in your Kubernetes cluster.
  • Access to the kube-state-metrics image hosted at registry.k8s.io/kube-state-metrics/kube-state-metrics:v2.14.0. You can pre-pull the image to your private registry.
  • The helm package manager. You run Helm commands in a Linux command-line terminal.

Install the Kubernetes operator with custom resource metrics enabled

You use the following variables in the AlloyDB Omni Kubernetes operator Helm chart to control the installation and removal of custom resource metrics:

Variable Type Description Default

crdMetrics.enabled

Boolean (true/false) Enables and disables custom resource metrics false (disable)

crdMetrics.ksmImage

String Full path of the ksmImage

registry.k8s.io/kube-state-metrics/kube-state-metrics:v2.14.0

If you haven't installed the operator Helm chart yet, set the variables when you run the Helm install.

To install Kubernetes operator with custom resource metrics enabled, follow these steps:

  1. Open a Linux command-line terminal.
  2. Define the following environment variables by running the following commands:

    export GCS_BUCKET=alloydb-omni-operator
    export HELM_PATH=$(gcloud storage cat gs://$GCS_BUCKET/latest)
    export OPERATOR_VERSION="${HELM_PATH%%/*}"
    
  3. Download the AlloyDB Omni Kubernetes operator:

    gcloud storage cp gs://$GCS_BUCKET/$HELM_PATH ./ --recursive
    
  4. If you pre-pull the public KSM image at registry.k8s.io/kube-state-metrics/kube-state-metrics:v2.14.0 to your private registry, skip this step and proceed to the next step. Otherwise, install the AlloyDB Omni Kubernetes operator with custom resource metrics, and then proceed to the final step in this procedure.

    helm install alloydbomni-operator alloydbomni-operator-${OPERATOR_VERSION}.tgz \
    --create-namespace \
    --namespace alloydb-omni-system \
    --set crdMetrics.enabled=true \
    --atomic \
    --timeout 5m
    
  5. Export the environment variable:

    export KSM_IMAGE=<full path to your KSM image>
    
  6. To install the AlloyDB Omni Kubernetes operator with custom resource metrics and your own KSM image location, run the following command:

    helm install alloydbomni-operator alloydbomni-operator-${OPERATOR_VERSION}.tgz \
    --create-namespace \
    --namespace alloydb-omni-system \
    --set crdMetrics.enabled=true,crdMetrics.ksmImage=${KSM_IMAGE} \
    --atomic \
    --timeout 5m
    

    A successful installation returns the following output:

    NAME: alloydbomni-operator
    LAST DEPLOYED: CURRENT_TIMESTAMP
    NAMESPACE: alloydb-omni-system
    STATUS: deployed
    REVISION: 1
    TEST SUITE: None
    
  7. Clean up the installation by deleting the downloaded AlloyDB Omni Kubernetes operator installation file. The file is named alloydbomni-operator-VERSION_NUMBER.tgz, and is located in your current working directory.

Enable custom resource metrics on an existing Kubernetes operator

If you already have the AlloyDB Omni Kubernetes operator version 1.3.0 (or later) installed, you can enable custom resource metrics by upgrading the existing operator.

  1. Open a Linux command-line terminal.
  2. Find your current AlloyDB Omni Kubernetes operator version and define the environment variables by running the following commands:

    export OPERATOR_VERSION=$(helm ls -n alloydb-omni-system  -o json | jq -r 'map(select(.name == "alloydbomni-operator") | .chart)| .[0]')
    
    OPERATOR_VERSION=${OPERATOR_VERSION#"alloydbomni-operator-"}
    
    echo $OPERATOR_VERSION
    
    export GCS_BUCKET=alloydb-omni-operator
    
    export HELM_PATH=$OPERATOR_VERSION/alloydbomni-operator-$OPERATOR_VERSION.tgz
    

    The output is your installed AlloyDB Omni Kubernetes operator version, for example, 1.3.0.

  3. Download the specified version operator:

    gcloud storage cp gs://$GCS_BUCKET/$HELM_PATH ./ --recursive
    
  4. Upgrade the AlloyDB Omni Kubernetes operator by running the following commands, and then go to the final step in this procedure. If you pre-pull the public KSM image at registry.k8s.io/kube-state-metrics/kube-state-metrics:v2.14.0 to your private registry, then skip this step and proceed directly to the next step.

    helm upgrade alloydbomni-operator alloydbomni-operator-${OPERATOR_VERSION}.tgz \
    --namespace alloydb-omni-system \
    --reuse-values \
    --set crdMetrics.enabled=true \
    --atomic \
    --timeout 5m
    
  5. Export the environment variable:

    export KSM_IMAGE=<full path to your KSM image>
    
  6. To perform the upgrade, run the following command:

    helm upgrade alloydbomni-operator alloydbomni-operator-${OPERATOR_VERSION}.tgz \
    --namespace alloydb-omni-system \
    --reuse-values \
    --set crdMetrics.enabled=true,crdMetrics.ksmImage=${KSM_IMAGE} \
    --atomic \
    --timeout 5m
    

    A successful upgrade returns the following output:

    NAME: alloydbomni-operator
    LAST DEPLOYED: CURRENT_TIMESTAMP
    NAMESPACE: alloydb-omni-system
    STATUS: deployed
    REVISION: 2
    TEST SUITE: None
    

Disable custom resource metrics on an existing Kubernetes operator

  1. Open a Linux command-line terminal.
  2. Run the following command:

    export OPERATOR_VERSION=$(helm ls -n alloydb-omni-system  -o json | jq -r 'map(select(.name == "alloydbomni-operator") | .chart)| .[0]')
    
    OPERATOR_VERSION=${OPERATOR_VERSION#"alloydbomni-operator-"}
    
  3. To disable custom resource metrics, run the following command:

    helm upgrade alloydbomni-operator alloydbomni-operator-${OPERATOR_VERSION}.tgz \
    --namespace alloydb-omni-system \
    --reuse-values \
    --set crdMetrics.enabled=false \
    --atomic \
    --timeout 5m
    

View metrics using the Prometheus API

To view your metrics, see View metrics using the Prometheus API.

What's next