Logs and metrics

This page explains how Google Distributed Cloud connected logs various types of information about its operation and how to view that information. The collection of some types of logs and metrics incur additional charges. For more information, see Billing for logs and metrics.

Configure logging and monitoring

Before you can start gathering logs and metrics, you must do the following:

  1. Enable the logging APIs by using the following commands:

    gcloud services enable opsconfigmonitoring.googleapis.com --project PROJECT_ID
    gcloud services enable logging.googleapis.com --project PROJECT_ID
    gcloud services enable monitoring.googleapis.com --project PROJECT_ID
    

    Replace PROJECT_ID with the ID of the target Google Cloud project.

  2. Grant the roles required to write logs and metrics:

    gcloud projects add-iam-policy-binding PROJECT_ID \
        --role roles/opsconfigmonitoring.resourceMetadata.writer \
        --member "serviceAccount:PROJECT_ID.svc.id.goog[kube-system/metadata-agent]"
    
    gcloud projects add-iam-policy-binding PROJECT_ID \
        --role roles/logging.logWriter \
         --member "serviceAccount:PROJECT_ID.svc.id.goog[kube-system/stackdriver-log-forwarder]"
    
    gcloud projects add-iam-policy-binding PROJECT_ID \
        --role roles/monitoring.metricWriter \
        --member "serviceAccount:PROJECT_ID.svc.id.goog[kube-system/gke-metrics-agent]"
    

    Replace PROJECT_ID with the ID of the target Google Cloud project.

Logs

This section lists the Cloud Logging resource types supported by Distributed Cloud. To view Distributed Cloud logs, use the Logs Explorer in the Google Cloud console. Distributed Cloud logging is always enabled.

The Distributed Cloud connected logged resource types are the following standard Kubernetes resources:

  • k8s_container
  • k8s_node

You can also capture and retrieve Distributed Cloud connected logs by using the Cloud Logging API. For information about how to configure this logging mechanism, see the documentation for Cloud Logging client libraries.

Metrics

This section lists the Cloud Monitoring metrics supported by Distributed Cloud. To view Distributed Cloud metrics, use the Metrics explorer in the Google Cloud console.

Distributed Cloud connected cluster metrics

For Distributed Cloud connected clusters, Distributed Cloud connected provides the following types of metrics generated by Distributed Cloud connected nodes:

  • Resource metrics provide information about Distributed Cloud connected node and Pod performance, such as CPU load and memory usage.
  • System application metrics provide information about Distributed Cloud connected system workloads, such as coredns.

For a list of these metrics, see Google Distributed Cloud and Google Distributed Cloud metrics.

Distributed Cloud connected does not provide metrics generated by the Kubernetes control planes associated with Distributed Cloud connected clusters.

Distributed Cloud connected hardware metrics

Distributed Cloud connected provides metrics for Distributed Cloud connected hardware by using the following resource types:

  • edgecontainer.googleapis.com/Machine
  • edgecontainer.googleapis.com/Rack

Machine resource metrics

Distributed Cloud connected writes the following Cloud Monitoring API metrics for the edgecontainer.googleapis.com/Machine resource:

Metric Description
/machine/cpu/total_cores
  • Kind: GAUGE
  • Type: INT
Total count of physical processor cores present on the machine.
/machine/cpu/usage_time
  • Kind: CUMULATIVE
  • Type: DOUBLE
  • Unit: Seconds
Cumulative CPU usage time for all cores on the machine. Type can be workload (customer workloads) or system (everything else).
/machine/cpu/utilization
  • Kind: GAUGE
  • Type: DOUBLE
CPU utilization percentage on the machine. Range is 0 to 1. Type can be workload (customer workloads) or system (everything else).
/machine/memory/total_bytes
  • Kind: GAUGE
  • Type: INT64
Byte count of total memory on the machine.
/machine/memory/used_bytes
  • Kind: GAUGE
  • Type: INT64
Byte count of used memory on the machine. memory_type is either evictable (reclaimable by the kernel) or non-evictable (not reclaimable).
/machine/memory/utilization
  • Kind: GAUGE
  • Type: DOUBLE
Memory utilization percentage on the machine. Range is 0 to 1. memory_type is either evictable (reclaimable by the kernel) or non-evictable (not reclaimable).
/machine/network/up
  • Kind: GAUGE
  • Type: BOOL
Indicates whether the network interface is up and running. Includes primary cards, secondary cards, and ports.
/machine/network/link_speed
  • Kind: GAUGE
  • Type: DOUBLE
  • Unit: Bytes per second
Link speed of the primary network interface card.
/machine/network/received_bytes_count
  • Kind: CUMULATIVE
  • Type: DOUBLE
Received byte count for the primary network interface card.
/machine/network/sent_bytes_count
  • Kind: CUMULATIVE
  • Type: DOUBLE
Sent byte count for the primary network interface card.
/machine/network/connectivity
  • Kind: GAUGE
  • Type: BOOL
Indicates whether the primary network interface card has internet connectivity.
/machine/disk/total_bytes
  • Kind: GAUGE
  • Type: INT64
Byte count of total disk space on the machine.
/machine/disk/used_bytes
  • Kind: GAUGE
  • Type: INT64
Byte count of used disk space on the machine.
/machine/disk/utilization
  • Kind: GAUGE
  • Type: DOUBLE
Disk space utilization percentage on the machine. Range is 0 to 1.
/machine/restart_count
  • Kind: CUMULATIVE
  • Type: INT
Number of restarts that the machine has undergone.
/machine/uptime
  • Kind: GAUGE
  • Type: INT
  • Unit: Seconds
Machine uptime since the last restart.
/machine/connected
  • Kind: GAUGE
  • Type: INT64
Indicates whether the machine is connected to Google Cloud.

Rack resource metrics

Distributed Cloud connected writes the following Cloud Monitoring API metrics for the edgecontainer.googleapis.com/Rack resource:

Metric Description
/router/bgp_up
  • Kind: GAUGE
  • Type: BOOL
Indicates whether the BGP peering session on the router is up and healthy. router_id identifies the specific router (up to 2 per rack).
/router/connected
  • Kind: GAUGE
  • Type: BOOL
Indicates whether the BGP router is connected to Google Cloud. router_id identifies the specific router (up to 2 per rack).

Export custom application logs and metrics

Distributed Cloud connected automatically exports logs for applications running on Distributed Cloud connected workloads. To export metrics for an application running on Distributed Cloud connected workloads, you must annotate it as described in the next section.

Annotate the workload to enable metrics export

To enable the collection of custom metrics from an application, add the following annotations to the application's Service or Deployment manifest:

  • prometheus.io/scrape: "true"
  • prometheus.io/path: "ENDPOINT_PATH". Replace ENDPOINT_PATH with the full path to the target application's metric endpoint.
  • prometheus.io/port: "PORT_NUMBER": the port on which the application's metric endpoint listens for connections.

Run an example application

In this section, you create an application that writes custom logs and exposes a custom metric endpoint.

  1. Save the following Service and Deployment manifests to a file named my-app.yaml. Notice that the Service has the annotation prometheus.io/scrape: "true":

    kind: Service
    apiVersion: v1
    metadata:
      name: "monitoring-example"
      namespace: "default"
      annotations:
        prometheus.io/scrape: "true"
    spec:
      selector:
        app: "monitoring-example"
      ports:
        - name: http
          port: 9090
    ---
    apiVersion: apps/v1
    kind: Deployment
    metadata:
      name: "monitoring-example"
      namespace: "default"
      labels:
        app: "monitoring-example"
    spec:
      replicas: 1
      selector:
        matchLabels:
          app: "monitoring-example"
      template:
        metadata:
          labels:
            app: "monitoring-example"
        spec:
          containers:
          - image: gcr.io/google-samples/prometheus-dummy-exporter:latest
            name: prometheus-example-exporter
            imagePullPolicy: Always
            command:
            - /bin/sh
            - -c
            - ./prometheus-dummy-exporter --metric-name=example_monitoring_up --metric-value=1 --port=9090
            resources:
              requests:
                cpu: 100m
    
  2. Create the Deployment and the Service:

    kubectl --kubeconfig apply -f my-app.yaml
    

View application logs

Console

  1. In the Google Cloud console, go to the Logs Explorer page.

    Go to Logs Explorer

  2. Click Resource.

  3. In the All resource types list, select Kubernetes Container.

  4. For Cluster name, select the name of your user cluster.

  5. For Namespace name, select default.

  6. Click Add, and then click Run query.

  7. In the Query results section, you can see log entries from the monitoring-example Deployment. For example:

    {
      "textPayload": "2020/11/14 01:24:24 Starting to listen on :9090\n",
      "insertId": "1oa4vhg3qfxidt",
      "resource": {
        "type": "k8s_container",
        "labels": {
          "pod_name": "monitoring-example-7685d96496-xqfsf",
          "cluster_name": ...,
          "namespace_name": "default",
          "project_id": ...,
          "location": "us-west1",
          "container_name": "prometheus-example-exporter"
        }
      },
      "timestamp": "2020-11-14T01:24:24.358600252Z",
      "labels": {
        "k8s-pod/pod-template-hash": "7685d96496",
        "k8s-pod/app": "monitoring-example"
      },
      "logName": "projects/.../logs/stdout",
      "receiveTimestamp": "2020-11-14T01:24:39.562864735Z"
    }
    

gcloud

  1. Use the gcloud logging read command:

    gcloud logging read 'resource.labels.project_id="PROJECT_ID" AND \
        resource.type="k8s_container" AND resource.labels.namespace_name="default"'
    

    Replace PROJECT_ID with the ID of your project.

  2. In the output, you can see log entries from the monitoring-example Deployment. For example:

    insertId: 1oa4vhg3qfxidt
    labels:
      k8s-pod/app: monitoring-example
      k8s- pod/pod-template-hash: 7685d96496
    logName: projects/.../logs/stdout
    receiveTimestamp: '2020-11-14T01:24:39.562864735Z'
    resource:
      labels:
        cluster_name: ...
        container_name: prometheus-example-exporter
        location: us-west1
        namespace_name: default
        pod_name: monitoring-example-7685d96496-xqfsf
        project_id: ...
      type: k8s_container
    textPayload: |
      2020/11/14 01:24:24 Starting to listen on :9090
    timestamp: '2020-11-14T01:24:24.358600252Z'
    

View application metrics

Your example application exposes a custom metric named example_monitoring_up. You can view the values of that metric in the Google Cloud console.

  1. In the Google Cloud console, go to the Metrics explorer page.

    Go to Metrics explorer

  2. For Resource type, select Kubernetes Pod.

  3. For Metric, select external/prometheus/example_monitoring_up.

  4. In the chart, you can see that example_monitoring_up has a repeated value of 1.

Collect metrics with Prometheus

Distributed Cloud connected supports the Prometheus metrics solution for collecting metrics on workloads running on clusters that have entered Survivability mode.

For this purpose, Distributed Cloud connected creates an unmanaged namespace with the name prom-monitoring when you create a Distributed Cloud connected cluster. We recommend that you use this namespace to deploy Prometheus. You can also copy the required resources from the prom-monitoring namespace to a namespace of your choice and deploy Prometheus there.

Collect Distributed Cloud connected metrics with Prometheus

To collect Distributed Cloud connected metrics with Prometheus, you must configure Prometheus metrics scraping. To do so, mount the prometheus-scrape-config ConfigMap in your Prometheus Pod and add the scrape configuration from the ConfigMap to your Prometheus configuration. For example:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: prometheus-deployment
  namespace: prom-monitoring
  labels:
    app: prom-monitoring
spec:
  replicas: 1
  selector:
    matchLabels:
      app: prom-monitoring
  template:
    metadata:
      labels:
        app: prom-monitoring
    spec:
      containers:
        - name: prometheus
          image: prom/prometheus:main
          args:
            - "--config.file=/etc/prometheus/prometheus.yml"
            - "--storage.tsdb.path=/prometheus/"
          ports:
            - containerPort: 9090
          volumeMounts:
            - name: prometheus-config-volume
              mountPath: /etc/prometheus/
            - name: prometheus-scrape-config-volume
              mountPath: /etc/prometheus/scrape/
            - name: prometheus-storage-volume
              mountPath: /prometheus/
            - name: prometheus-scrape-token
              mountPath: /var/run/secrets/kubernetes.io/stackdriver-prometheus-scrape
            - name: stackdriver-prometheus-scrape-cert
              mountPath: /certs/stackdriver-prometheus-scrape
            - name: metrics-providers-ca
              mountPath: /certs/metrics-providers
            - name: stackdriver-prometheus-etcd-scrape
              mountPath: /stackdriver-prometheus-etcd-scrape
      volumes:
        - name: prometheus-storage-volume
          emptyDir: {}
        - name: prometheus-config-volume
          configMap:
            defaultMode: 420
            name: prometheus-config
        - name: prometheus-scrape-config-volume
          configMap:
            defaultMode: 420
            name: prometheus-scrape-config
        - name: prometheus-scrape-token
          secret:
            defaultMode: 420
            secretName: prometheus-scrape
        - name: stackdriver-prometheus-scrape-cert
          secret:
            defaultMode: 420
            secretName: stackdriver-prometheus-scrape-cert
        - name: metrics-providers-ca
          secret:
            defaultMode: 420
            items:
            - key: ca.crt
              path: ca.crt
            secretName: metrics-providers-ca
        - name: stackdriver-prometheus-etcd-scrape
          secret:
            defaultMode: 420
            optional: true
            secretName: stackdriver-prometheus-etcd-scrape

To collect workload metrics with Prometheus, you must add annotations to the Services and Pods executing the target workloads as follows:

  • To send metrics to both Cloud Monitoring and Prometheus, use the annotations described in Export custom logs and metrics.

  • To send metrics only to Prometheus, use the following annotations:

prometheus.io/unmanaged_scrape: "true"
prometheus.io/unmanaged_path: "ENDPOINT_PATH"
prometheus.io/unmanaged_port: "PORT_NUMBER"

Collect metrics with Kafka

Distributed Cloud connected supports the Apache Kafka solution for collecting metrics on workloads running on your Distributed Cloud connected deployment.

You must have a functional Kafka deployment before completing the steps in this section. Your Kafka brokers must advertise their listeners in PLAINTEXT mode. SSL/SASL-related variables are not supported.

To configure a cluster for Kafka logging, you must create a JSON file that configures add-ons that you want to run on the cluster. Specify this file when creating a cluster using the --system-addons-config flag. If you need to modify the Kafka configuration, you must delete and re-create the cluster with the new Kafka settings.

Add the following section to the system add-ons configuration file:

{
        "systemAddonsConfig": {
          "unmanagedKafkaConfig": {
            "brokers": "BROKERS",
            "topics": "TOPICS"
     "topic_key": "TOPIC_KEY, 
          }
      }
}

Replace the following:

  • BROKERS: a comma-separated list of broker IP address and port pairs in ip_address:port format.
  • TOPICS: a comma-separated list Kafka topics.
  • TOPIC_KEY: a Kafka topic key; this allows Kafka to select a topic if multiple topics exist.

If Kafka is not collecting metrics after you've created your Distributed Cloud connected cluster, check the following:

  • Server side: Check the error logs on your Kafka deployment for indications of a problem.
  • Client side: Contact Google Support to retrieve and examine system Pod logs.

What's next