Parca Operator

The Parca Operator can be used to deploy and operator Parca. It can be used to deploy the Parca Server and Parca Agent. Besides that the Parca Operator can also be used to configure the pull-based ingestion via Custom Resource Definitions.

Note

The Parca Operator is work in progress. While I plan to add support for the deployment of the Parca Server and Parca Agent, these features are not implemented yet.

This means, that the Parca Operator can currently only be used to configure the pull-based ingestion via the ParcaScrapeConfig Custom Resource Definition.

Installation

The Parca Operator can be installed via Helm:

helm upgrade --install parca-operator oci://ghcr.io/ricoberger/charts/parca-operator --version <VERSION>

Make sure that you set the following environment variables for the Parca Operator:

• PARCA_CONFIG_SOURCE: The path to the configration file for Parca. This file is used as source for the generated Parca configration and should contain the object_storage configuration.
• PARCA_CONFIG_TARGET_NAME and PARCA_CONFIG_TARGET_NAMESPACE: The name and namespace of secret which should be generated. The secret contains a parca.yaml key with the generated configuration for Parca. The generated configuration file contains the content of the source configuration and the configuration for all scrape configuration created via a ParcaScrapeConfig resource.

API Reference

ParcaScrapeConfig

apiVersion: parca.ricoberger.de/v1alpha1
kind: ParcaScrapeConfig
metadata:
  name:
  namespace:
spec:
  # Selector is the selector for the Pods which should be scraped by Parca.
  selector:
    matchLabels:
    matchExpressions:
  # ScrapeConfig is the scrape configuration as it can be set in the Parca
  # configuration.
  scrapeConfig:
    # Job is the job name of the section in the configurtion. If no job name is
    # provided, it will be automatically generated based on the name and
    # namespace of the CR: "namespace/name"
    job:
    # Port is the name of the port of the Pods which is used to expose the
    # profiling endpoints.
    port:
    # PortNumber is the number of the port which is used to expose the
    # profiling endpoints. This can be used instead of the port field. If the
    # port is not named.
    portNumber:
    # Params is a set of query parameters with which the target is scraped.
    params:
    # Interval defines how frequently to scrape the targets of this scrape
    # config.
    interval:
    # Timeout defines the timeout for scraping targets of this config.
    timeout:
    # Schema sets the URL scheme with which to fetch metrics from targets.
    scheme:
    # ProfilingConfig defines the profiling config for the targets, see
    # https://www.parca.dev/docs/ingestion#pull-based for more information.
    profiling_config:
    # RelabelConfigs allows dynamic rewriting of the label set for the targets,
    # see https://prometheus.io/docs/prometheus/latest/configuration/configuration/#relabel_config
    # for more information.
    relabel_configs:

Example

apiVersion: parca.ricoberger.de/v1alpha1
kind: ParcaScrapeConfig
metadata:
  name: parca-server
  namespace: parca
spec:
  selector:
    matchLabels:
      app: parca-server
  scrapeConfig:
    port: grpc
    interval: 45s
    timeout: 60s
    profilingConfig:
      pprofConfig:
        fgprof:
          enabled: true
          path: /debug/pprof/fgprof

Development

After modifying the *_types.go file always run the following command to update the generated code for that resource type:

make generate

The above Makefile target will invoke the controller-gen utility to update the api/v1alpha1/zz_generated.deepcopy.go file to ensure our API's Go type definitons implement the runtime.Object interface that all Kind types must implement.

Once the API is defined with spec/status fields and CRD validation markers, the CRD manifests can be generated and updated with the following command:

make manifests

This Makefile target will invoke controller-gen to generate the CRD manifests at charts/parca-operator/crds/parca.ricoberger.de_<CRD>.yaml.

Deploy the CRD and run the operator locally with the default Kubernetes config file present at $HOME/.kube/config:

export PARCA_CONFIG_SOURCE=parca.yaml
export PARCA_CONFIG_TARGET_NAME=parca-generated
export PARCA_CONFIG_TARGET_NAMESPACE=parca

make run