Documentation
¶
Index ¶
Constants ¶
View Source
const ( // MetricLabelErrorDefaultVal is the default string value that represents "error type unknown" MetricLabelErrorDefaultVal = "" // Well-known metricLabelError values NodeClaimNotFoundError = "NodeClaimNotFoundError" NodeClassNotReadyError = "NodeClassNotReadyError" InsufficientCapacityError = "InsufficientCapacityError" )
Variables ¶
View Source
var ( ErrorsTotal = opmetrics.NewPrometheusCounter( crmetrics.Registry, prometheus.CounterOpts{ Namespace: metrics.Namespace, Subsystem: "cloudprovider", Name: "errors_total", Help: "Total number of errors returned from CloudProvider calls.", }, []string{ metricLabelController, metricLabelMethod, metricLabelProvider, metricLabelError, }, ) )
View Source
var MethodDuration = opmetrics.NewPrometheusHistogram( crmetrics.Registry, prometheus.HistogramOpts{ Namespace: metrics.Namespace, Subsystem: "cloudprovider", Name: "duration_seconds", Help: "Duration of cloud provider method calls. Labeled by the controller, method name and provider.", }, []string{ metricLabelController, metricLabelMethod, metricLabelProvider, }, )
Functions ¶
func Decorate ¶
func Decorate(cloudProvider cloudprovider.CloudProvider) cloudprovider.CloudProvider
Decorate returns a new `CloudProvider` instance that will delegate all method calls to the argument, `cloudProvider`, and publish aggregated latency metrics. The value used for the metric label, "controller", is taken from the `Context` object passed to the methods of `CloudProvider`.
Do not decorate a `CloudProvider` multiple times or published metrics will contain duplicated method call counts and latencies.
func GetErrorTypeLabelValue ¶
GetErrorTypeLabelValue is a convenience func that returns a string representation of well-known CloudProvider error types
Types ¶
This section is empty.
Click to show internal directories.
Click to hide internal directories.