Important: This documentation is about an older version. It's relevant only to the release noted, many of the features and functions have been updated or replaced. Please view the current version.

Query-frontend to work with Prometheus

Open source

Configure the Grafana Mimir query-frontend to work with Prometheus

You can use the Grafana Mimir query-frontend with any Prometheus-API compatible service, including Prometheus and Thanos. Use this configuration file to benefit from query parallelization and caching.

# You can use the Mimir query frontend with any Prometheus-API compatible
# service, including Prometheus and Thanos.  Use this config file to get
# the benefits of query parallelisation and caching.

# Disable the requirement that every request to Mimir has a
# X-Scope-OrgID header. `anonymous` will be substituted in instead.
multitenancy_enabled: false

# We only want to run the query-frontend module.
target: query-frontend

# We don't want the default /api/prom or /prometheus prefixes on endpoints.
api:
  prometheus_http_prefix: ''

server:
  http_listen_port: 9091

frontend:
  split_queries_by_interval: 24h
  align_queries_with_step: true
  cache_results: true

  results_cache:
    backend: "memcached"

    memcached:
      # You can either configure a headless service in Kubernetes and Mimir will discover the individual
      # instances using a SRV DNS query (host) or list comma separated memcached addresses.
      addresses: "dnssrvnoa+memcached.mimir.svc.cluster.local:11211"

  log_queries_longer_than: 1s

  # The Prometheus URL to which the query-frontend should connect to.
  downstream_url: http://prometheus.mydomain.com