Documentationbreadcrumb arrow Grafana Cloudbreadcrumb arrow Observe and actbreadcrumb arrow Instrument and send databreadcrumb arrow Grafana Alloybreadcrumb arrow Configurebreadcrumb arrow Clustering
Grafana Cloud

Configure clustering

You can configure Alloy to run with clustering so that individual Alloy instances can work together for workload distribution and high availability.

To configure clustering, complete these two steps:

  1. Enable clustering in Alloy: Configure Alloy itself to join a cluster using Helm chart settings or command-line flags.
  2. Enable clustering in components: Add a clustering block to each component that should participate in workload distribution.

Note

Cluster mode at the Alloy level doesn’t automatically distribute workload. You must also enable clustering in each component that should participate in the cluster. Only components with a configured clustering block distribute workload.

Enable clustering in components

Several components support workload distribution through clustering, including prometheus.scrape, loki.source.kubernetes, and pyroscope.scrape. Refer to Distribute workload across cluster nodes for the complete list.

To enable clustering in a component, add a clustering block with enabled = true:

Alloy 
prometheus.scrape "example" {
  targets    = discovery.kubernetes.pods.targets
  forward_to = [prometheus.remote_write.default.receiver]

  clustering {
    enabled = true
  }
}

Next steps