Documentation
¶
Overview ¶
Package metrics provides the framework's Prometheus implementation of component.MetricsRecorder.
It records two things: the status condition metrics of go-crd-condition-metrics, and the framework's own resource-level apply counters. Wire one Recorder per controller and share a single Collectors across the process:
var (
conditions = ocm.NewOperatorConditionsGauge("myoperator")
collectors = metrics.NewCollectors()
)
func init() {
ctrlmetrics.Registry.MustRegister(conditions, collectors)
}
rec := component.ReconcileContext{
// ...
Metrics: metrics.NewRecorder("webapp", conditions, collectors),
}
The controller name must match the controller-runtime controller name (the lower-cased kind passed to For, unless Named overrides it) so the dashboards and alerts shipped under observability/ correlate the framework's series with controller-runtime's.
Index ¶
- type Collectors
- type Recorder
- func (r *Recorder) RecordConditionFor(kind string, object ocm.ObjectLike, ...)
- func (r *Recorder) RecordResourceApply(labels component.ResourceMetricLabels, operation concepts.ConvergingOperation)
- func (r *Recorder) RecordResourceApplyError(labels component.ResourceMetricLabels)
- func (r *Recorder) RemoveConditionsFor(kind string, object ocm.ObjectLike) int
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Collectors ¶
type Collectors struct {
// contains filtered or unexported fields
}
Collectors holds the framework's resource-level Prometheus collectors: ocf_resource_apply_total and ocf_resource_apply_errors_total.
Construct one per process and register it once, before any Recorder that uses it starts reconciling:
collectors := metrics.NewCollectors() ctrlmetrics.Registry.MustRegister(collectors)
Every controller in the process then shares it, told apart by the `controller` label their Recorder supplies.
The series are keyed only by the operator's static topology: controller, owner kind, component, resource identifier, kind and operation. No owner name or namespace appears, so the same handful of series covers three owners or three thousand, and no series needs removing when an owner is deleted. That holds only while every resource identifier stays low-cardinality, which is the contract WithMetricsIdentifier documents on the resource builders.
func NewCollectors ¶
func NewCollectors() *Collectors
NewCollectors creates the framework's resource-level collectors. Register the result with a Prometheus registry, typically controller-runtime's.
func (*Collectors) Collect ¶
func (c *Collectors) Collect(ch chan<- prometheus.Metric)
Collect implements prometheus.Collector.
func (*Collectors) Describe ¶
func (c *Collectors) Describe(ch chan<- *prometheus.Desc)
Describe implements prometheus.Collector.
type Recorder ¶
type Recorder struct {
// contains filtered or unexported fields
}
Recorder is the framework's Prometheus implementation of component.MetricsRecorder. Create one per controller with NewRecorder and assign it to component.ReconcileContext.Metrics.
Both halves are independently optional: a Recorder built without a conditions gauge records only resource metrics, and one built without collectors records only condition metrics.
func NewRecorder ¶
func NewRecorder( controller string, conditions *ocm.OperatorConditionsGauge, collectors *Collectors, ) *Recorder
NewRecorder creates a Recorder for the named controller.
controller is the value of the `controller` label on every series the recorder emits, condition metrics included, so it must match the name controller-runtime uses for that controller (the lower-cased kind passed to For, unless Named overrides it); the shipped dashboards and alerts filter both families with one controller label. conditions and collectors are the shared, registered collectors; passing nil for either disables that half of the recording rather than panicking at reconcile time.
func (*Recorder) RecordConditionFor ¶
func (r *Recorder) RecordConditionFor( kind string, object ocm.ObjectLike, conditionType, conditionStatus, conditionReason string, lastTransitionTime time.Time, extraLabelValues ...string, )
RecordConditionFor records a condition change for the given object and kind. It is a no-op when the recorder was built without a conditions gauge.
func (*Recorder) RecordResourceApply ¶
func (r *Recorder) RecordResourceApply( labels component.ResourceMetricLabels, operation concepts.ConvergingOperation, )
RecordResourceApply records one framework apply of a managed resource. It is a no-op when the recorder was built without collectors.
func (*Recorder) RecordResourceApplyError ¶
func (r *Recorder) RecordResourceApplyError(labels component.ResourceMetricLabels)
RecordResourceApplyError records one failed framework apply of a managed resource. It is a no-op when the recorder was built without collectors.
func (*Recorder) RemoveConditionsFor ¶
func (r *Recorder) RemoveConditionsFor(kind string, object ocm.ObjectLike) int
RemoveConditionsFor deletes every condition metric for the given object, returning the number of time series removed. Call it when the object is deleted, so its condition series do not outlive it.
There is deliberately no counterpart for the resource-level metrics: those series carry no owner identity, so they do not accumulate per object, and deleting a counter mid-flight reads downstream as a counter reset.
It returns zero when the recorder was built without a conditions gauge.
Source Files