This repository contains an implementation of the Kubernetes custom metrics API (custom-metrics.metrics.k8s.io/v1alpha1), suitable for use with the autoscaling/v2 Horizontal Pod Autoscaler in Kubernetes 1.6+.
The adapter takes the standard Kubernetes generic API server arguments (including those for authentication and authorization). By default, it will attempt to using Kubernetes in-cluster config to connect to the cluster.
It takes the following addition arguments specific to configuring how the adapter talks to Prometheus and the main Kubernetes cluster:
-
--lister-kubeconfig=<path-to-kubeconfig>: This configures how the adapter talks to a Kubernetes API server in order to list objects when operating with label selectors. By default, it will use in-cluster config. -
--metrics-relist-interval=<duration>: This is the interval at which to update the cache of available metrics from Prometheus. -
--rate-interval=<duration>: This is the duration used when requesting rate metrics from Prometheus. It must be larger than your Prometheus collection interval. -
--prometheus-url=<url>: This is the URL used to connect to Prometheus. It will eventually contain query parameters to configure the connection.
The adapter gathers the names of available metrics from Prometheus at the specified interval. Only metrics of the following forms are considered:
-
"container" metrics (cAdvisor container metrics): series with a name starting with
container_, as well as non-emptynamespaceandpod_namelabels. -
"namespaced" metrics (metrics describing namespaced Kubernetes objects): series with non-empty namespace labels (which don't start with
container_).
Note: Currently, metrics on non-namespaced objects (besides namespaces themselves) are not supported.
Metrics in Prometheus are converted in the custom-metrics-API metrics as follows:
-
The metric name and type are decided:
- For container metrics, the
container_prefix is removed - If the metric has the
_totalsuffix, it is marked as a counter metric, and the suffix is removed - If the metric has the
_seconds_totalsuffix, it is marked as a seconds counter metric, and the suffix is removed. - If the metric has none of the above suffixes, is is marked as a gauge metric, and the metric name is used as-is
- For container metrics, the
-
Relevant resources are associated with the metric:
- container metrics are associated with pods only
- for non-container metrics, each label on the series is considered. If that label represents a resource (without the group) available on the server, the metric is associated with that resource. A metric may be associated with multiple resources.
When retrieving counter and seconds-counter metrics, the adapter requests the metrics as a rate over the configured amount of time. For metrics with multiple associated resources, the adapter requests the metric aggregated over all non-requested metrics.
The adapter does not consider resources consumed by the "POD" container, which exists as part of all Kubernetes pods running in Docker simply supports the existance of the pod's shared network namespace.