Bringing your own Prometheus

Even though Linkerd comes with its own Prometheus instance, there can be cases where using an external instance makes more sense for various reasons.

Note

Note that this approach requires you to manually add and maintain additional scrape configuration in your Prometheus configuration. If you prefer to use the default Linkerd Prometheus add-on, you can export the metrics to your existing monitoring infrastructure following these instructions.

This tutorial shows how to configure an external Prometheus instance to scrape both the control plane as well as the proxy’s metrics in a format that is consumable both by a user as well as Linkerd control plane components like web, etc.

There are two important points to tackle here.

Configuring external Prometheus instance to get the Linkerd metrics.
Configuring the Linkerd control plane components to use that Prometheus.

Prometheus Scrape Configuration

The following scrape configuration has to be applied to the external Prometheus instance.

Note

The below scrape configuration is a subset of linkerd-prometheus scrape configuration.

Before applying, it is important to replace templated values (present in {{}}) with direct values for the below configuration to work.

    - job_name: 'linkerd-controller'

      scrape_interval: 10s
      scrape_timeout: 10s

      kubernetes_sd_configs:
      - role: pod
        namespaces:
          names: ['{{.Values.global.namespace}}']
      relabel_configs:
      - source_labels:
        - __meta_kubernetes_pod_label_linkerd_io_control_plane_component
        - __meta_kubernetes_pod_container_port_name
        action: keep
        regex: (.*);admin-http$
      - source_labels: [__meta_kubernetes_pod_container_name]
        action: replace
        target_label: component

    - job_name: 'linkerd-service-mirror'

      scrape_interval: 10s
      scrape_timeout: 10s

      kubernetes_sd_configs:
      - role: pod
      relabel_configs:
      - source_labels:
        - __meta_kubernetes_pod_label_linkerd_io_control_plane_component
        - __meta_kubernetes_pod_container_port_name
        action: keep
        regex: linkerd-service-mirror;admin-http$
      - source_labels: [__meta_kubernetes_pod_container_name]
        action: replace
        target_label: component

    - job_name: 'linkerd-proxy'

      scrape_interval: 10s
      scrape_timeout: 10s

      kubernetes_sd_configs:
      - role: pod
      relabel_configs:
      - source_labels:
        - __meta_kubernetes_pod_container_name
        - __meta_kubernetes_pod_container_port_name
        - __meta_kubernetes_pod_label_linkerd_io_control_plane_ns
        action: keep
        regex: ^{{default .Values.global.proxyContainerName "linkerd-proxy" .Values.global.proxyContainerName}};linkerd-admin;{{.Values.global.namespace}}$
      - source_labels: [__meta_kubernetes_namespace]
        action: replace
        target_label: namespace
      - source_labels: [__meta_kubernetes_pod_name]
        action: replace
        target_label: pod
      # special case k8s' "job" label, to not interfere with prometheus' "job"
      # label
      # __meta_kubernetes_pod_label_linkerd_io_proxy_job=foo =>
      # k8s_job=foo
      - source_labels: [__meta_kubernetes_pod_label_linkerd_io_proxy_job]
        action: replace
        target_label: k8s_job
      # drop __meta_kubernetes_pod_label_linkerd_io_proxy_job
      - action: labeldrop
        regex: __meta_kubernetes_pod_label_linkerd_io_proxy_job
      # __meta_kubernetes_pod_label_linkerd_io_proxy_deployment=foo =>
      # deployment=foo
      - action: labelmap
        regex: __meta_kubernetes_pod_label_linkerd_io_proxy_(.+)
      # drop all labels that we just made copies of in the previous labelmap
      - action: labeldrop
        regex: __meta_kubernetes_pod_label_linkerd_io_proxy_(.+)
      # __meta_kubernetes_pod_label_linkerd_io_foo=bar =>
      # foo=bar
      - action: labelmap
        regex: __meta_kubernetes_pod_label_linkerd_io_(.+)
      # Copy all pod labels to tmp labels
      - action: labelmap
        regex: __meta_kubernetes_pod_label_(.+)
        replacement: __tmp_pod_label_$1
      # Take `linkerd_io_` prefixed labels and copy them without the prefix
      - action: labelmap
        regex: __tmp_pod_label_linkerd_io_(.+)
        replacement:  __tmp_pod_label_$1
      # Drop the `linkerd_io_` originals
      - action: labeldrop
        regex: __tmp_pod_label_linkerd_io_(.+)
      # Copy tmp labels into real labels
      - action: labelmap
        regex: __tmp_pod_label_(.+)

The running configuration of the builtin prometheus can be used as a reference.

kubectl -n linkerd  get configmap linkerd-prometheus-config -o yaml

Control Plane Components Configuration

Linkerd’s control plane components like public-api, etc depend on the Prometheus instance to power the dashboard and CLI.

The global.prometheusUrl field gives you a single place through which all these components can be configured to an external Prometheus URL. This is allowed both through the CLI and Helm.

CLI

This can be done by passing a file with the above field to the config flag, which is available both through linkerd install and linkerd upgrade commands

global:
  prometheusUrl: existing-prometheus.xyz:9090

Once applied, this configuration is persistent across upgrades, without having the user passing it again. The same can be overwritten as needed.

When using an external Prometheus and configuring the global.prometheusUrl field, Linkerd’s Prometheus will still be included in installation.

If you wish to disable this included Prometheus, be sure to include the following configuration as well:

prometheus:
  enabled: false

Helm

The same configuration can be applied through values.yaml when using Helm. Once applied, Helm makes sure that the configuration is persistent across upgrades.

More information on installation through Helm can be found here