Skip to content

Latest commit

 

History

History
122 lines (79 loc) · 6.37 KB

model.md

File metadata and controls

122 lines (79 loc) · 6.37 KB

Heapster Metric Model

The Heapster Model is a structured representation of metrics for Kubernetes clusters, which is exposed through a set of REST API endpoints. It allows the extraction of up to 1 hour of historical data for any Container, Pod, Node or Namespace in the cluster, as well as the cluster itself. The model can also export the Average, Max and 95th Percentile for each one of these entities, over a duration of 1 minute, 1 hour or 24 hours.

NOTICE: this document represents the state of the model at the next Heapster release, and is not representative of the current state of the master branch TODO(afein): remove NOTICE prior to next release

Usage

The Heapster Model can be enabled by initializing Heapster with the -use_model=true flag. The resolution of the model can be configured through the -model_resolution flag, which will cause the model to store historical data at the specified resolution. If the model_resolution flag is not specified, the default resolution of 1 minute will be used.

API documentation

A detailed documentation of each API endpoint is listed below.

All endpoints ending in /metrics/{metric-name} can accept the optional start and end query parameters that represent the start and end time of the requested timeseries. The result will be a list of (Timestamp, Value) pairs in the time range (start, end]. start and end are strings formatted according to RFC3339. If start is not defined, it is assumed as the zero Unix epoch time. If end is not defined, then all data later than start will be returned.

Cluster-level Metrics and Stats

/api/v1/model/: The root path of the model API, displays all browsable subpaths.

/api/v1/model/metrics/: Returns a list of available cluster-level metrics.

/api/v1/model/metrics/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested cluster-level metric, between the time range specified by start and end.

/api/v1/model/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each cluster metric.

Node-level Metrics and Stats

/api/v1/model/nodes: Returns a list of all available nodes, along with their latest CPU and Memory Usage values.

/api/v1/model/nodes/{node-name}: Returns all browsable subpaths for a specific node.

/api/v1/model/nodes/{node-name}/pods: Returns a list of all available pods under a given node, along with their latest CPU and Memory Usage values

/api/v1/model/nodes/{node-name}/metrics/: Returns a list of available node-level metrics

/api/v1/model/metrics/nodes/{node-name}/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested node-level metric, within the time range specified by start and end.

/api/v1/model/nodes/{node-name}/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each node metric

Namespace-level Metrics and Stats

/api/v1/model/namespaces: Returns a list of all available namespaces, along with their latest CPU and Memory Usage values.

/api/v1/model/namespaces/{namespace-name}: Returns all browsable subpaths for a specific namespace.

/api/v1/model/namespaces/{namespace-name}/metrics/: Returns a list of available namespace-level metrics

/api/v1/model/namespaces/{namespace-name}/metrics/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested namespace-level metric, within the time range specified by start and end.

/api/v1/model/namespaces/{namespace-name}/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each namespace metric.

Pod-level Metrics and Stats

/api/v1/model/namespaces/{namespace-name}/pods: Returns a list of all available pods under a given namespace, along with their latest CPU and Memory Usage values.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}: Returns all browsable subpaths for a specific Pod.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/metrics/: Returns a list of available pod-level metrics

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/metrics/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested pod-level metric, within the time range specified by start and end.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each pod-level metric.

Container-level Metrics and Stats

Container metrics and stats are accessible for both containers that belong to pods, as well as for free containers running in each node.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/containers: Returns a list of all available containers under a given pod, along with their latest CPU and Memory Usage values.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/containers/{container-name}: Returns all browsable subpaths for a specific container.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/containers/{container-name}/metrics/: Returns a list of available container-level metrics

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/containers/{container-name}/metrics/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested container-level metric, within the time range specified by start and end.

/api/v1/model/namespaces/{namespace-name}/pods/{pod-name}/containers/{container-name}/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each container metric.

/api/v1/model/nodes/{node-name}/freecontainers: Returns a list of all available free containers under a given node, along with their latest CPU and Memory Usage values.

/api/v1/model/nodes/{node-name}/freecontainers/{container-name}: Returns all browsable subpaths for a specific free container.

/api/v1/model/nodes/{node-name}/freecontainers/{container-name}/metrics/: Returns a list of available container-level metrics

/api/v1/model/nodes/{node-name}/freecontainers/{container-name}/metrics/{metric-name}?start=X&end=Y: Returns a set of (Timestamp, Value) pairs for the requested container-level metric, within the time range specified by start and end.

/api/v1/model/nodes/{node-name}/freecontainers/{container-name}/stats/: Exposes the average, max and 95th percentile over the past minute, hour and day for each container metric.