docs/sample-config.yaml

rules:
# Each rule represents a some naming and discovery logic.
# Each rule is executed independently of the others, so
# take care to avoid overlap.  As an optimization, rules
# with the same `seriesQuery` but different
# `name` or `seriesFilters` will use only one query to
# Prometheus for discovery.

# some of these rules are taken from the "default" configuration, which
# can be found in pkg/config/default.go

# this rule matches cumulative cAdvisor metrics measured in seconds
- seriesQuery: '{__name__=~"^container_.*",container!="POD",namespace!="",pod!=""}'
  resources:
    # skip specifying generic resource<->label mappings, and just
    # attach only pod and namespace resources by mapping label names to group-resources
    overrides:
      namespace: {resource: "namespace"}
      pod: {resource: "pod"}
  # specify that the `container_` and `_seconds_total` suffixes should be removed.
  # this also introduces an implicit filter on metric family names
  name:
    # we use the value of the capture group implicitly as the API name
    # we could also explicitly write `as: "$1"`
    matches: "^container_(.*)_seconds_total$"
  # specify how to construct a query to fetch samples for a given series
  # This is a Go template where the `.Series` and `.LabelMatchers` string values
  # are available, and the delimiters are `<<` and `>>` to avoid conflicts with
  # the prometheus query language
  metricsQuery: "sum(rate(<<.Series>>{<<.LabelMatchers>>,container!="POD"}[2m])) by (<<.GroupBy>>)"

# this rule matches cumulative cAdvisor metrics not measured in seconds
- seriesQuery: '{__name__=~"^container_.*_total",container!="POD",namespace!="",pod!=""}'
  resources:
    overrides:
      namespace: {resource: "namespace"}
      pod: {resource: "pod"}
  seriesFilters:
  # since this is a superset of the query above, we introduce an additional filter here
  - isNot: "^container_.*_seconds_total$"
  name: {matches: "^container_(.*)_total$"}
  metricsQuery: "sum(rate(<<.Series>>{<<.LabelMatchers>>,container!="POD"}[2m])) by (<<.GroupBy>>)"

# this rule matches cumulative non-cAdvisor metrics
- seriesQuery: '{namespace!="",__name__!="^container_.*"}'
  name: {matches: "^(.*)_total$"}
  resources:
    # specify an a generic mapping between resources and labels.  This
    # is a template, like the `metricsQuery` template, except with the `.Group`
    # and `.Resource` strings available.  It will also be used to match labels,
    # so avoid using template functions which truncate the group or resource.
    # Group will be converted to a form acceptible for use as a label automatically.
    template: "<<.Resource>>"
    # if we wanted to, we could also specify overrides here
  metricsQuery: "sum(rate(<<.Series>>{<<.LabelMatchers>>,container!="POD"}[2m])) by (<<.GroupBy>>)"

# this rule matches only a single metric, explicitly naming it something else
# It's series query *must* return only a single metric family
- seriesQuery: 'cheddar{sharp="true"}'
  # this metric will appear as "cheesy_goodness" in the custom metrics API
  name: {as: "cheesy_goodness"}
  resources:
    overrides:
      # this should still resolve in our cluster
      brand: {group: "cheese.io", resource: "brand"}
  metricsQuery: 'count(cheddar{sharp="true"})'

# external rules are not tied to a Kubernetes resource and can reference any metric
# https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale-walkthrough/#autoscaling-on-metrics-not-related-to-kubernetes-objects
externalRules:
- seriesQuery: '{__name__="queue_consumer_lag",name!=""}'
  metricsQuery: sum(<<.Series>>{<<.LabelMatchers>>}) by (name)
- seriesQuery: '{__name__="queue_depth",topic!=""}'
  metricsQuery: sum(<<.Series>>{<<.LabelMatchers>>}) by (name)
  # Kubernetes metric queries include a namespace in the query by default
  # but you can explicitly disable namespaces if needed with "namespaced: false"
  # this is useful if you have an HPA with an external metric in namespace A
  # but want to query for metrics from namespace B
  resources:
    namespaced: false

# TODO: should we be able to map to a constant instance of a resource
# (e.g. `resources: {constant: [{resource: "namespace", name: "kube-system"}}]`)?