metrics

type ClusterRecorder interface {
	// cluster/addons_expected
	// Entity: k8s_cluster
	// Label: name / Value: (name of addon)
	// Label: version / Value: (version of addon)
	// Label: controller / Value: (DaemonSet or Deployment)
	// Value: count of addons
	RecordAddonCounts(ctx context.Context, counts []LabelCount)

	// cluster/node_available (SLI)
	// Entity: k8s_cluster
	// Label: nodepool / Values: (set of nodepools)
	// Label: zone / Values: (set of zones)
	// Label: available / Value (True/False)
	// Label: ready / Value (True/False)
	// Label: schedulable / Value (True/False)
	// Label: done_warming / Value (True/False)
	// Value: count of available nodes
	RecordNodeAvailabilities(ctx context.Context, counts []LabelCount)

	// cluster/node_condition
	// Entity: k8s_cluster
	// Label: nodepool / Values: (set of nodepools)
	// Label: zone / Values: (set of zones)
	// Label: type / Values: (set of conditions)
	// Label: status / Values: (True/False/Unknown)
	// Value: count of nodes reflecting condition
	RecordNodeConditions(ctx context.Context, counts []LabelCount)
}

type LabelCount struct {
	Labels map[string]string
	Count  int
}

type NodeRecorder interface {
	// addon/available
	// Entity: k8s_node
	// Label: nodepool / Value (name of nodepool)
	// Label: zone / Value (name of zone)
	// Label: name / Value: (name of addon)
	// Label: version / Value: (version of addon)
	// Label: controller / Value: (DaemonSet or Deployment)
	// Label: available / Value (True/False)
	// Label: node_available / Value (True/False)
	// Label: running / Value: (True/False)
	// Label: stable / Value: (True/False)
	// Label: healthy / Value: (True/False/Unknown/Error)
	// Value: 1
	RecordAddonAvailabilies(ctx context.Context, counts []LabelCount)

	// addon/restart
	// Entity: k8s_node
	// Label: nodepool / Value (name of nodepool)
	// Label: zone / Value (name of zone)
	// Label: name / Value: (name of addon)
	// Label: version / Value: (version of addon)
	// Label: controller / Value: (DaemonSet or Deployment)
	// Label: container_name / Value: (name of container in pod)
	// Label: reason / Value: (reason for restart)
	// Label: exit_code / Value: (exit code)
	// Value: 1
	RecordContainerRestart(ctx context.Context, labels map[string]string)

	// addon/control_plane_available
	// Entity: k8s_node
	// Label: nodepool / Value (name of nodepool)
	// Label: zone / Value (name of zone)
	// Label: available / Value (True/False)
	// Value: 1
	RecordAddonControlPlaneAvailability(ctx context.Context, labels map[string]string)

	// node/node_condition
	// Entity: k8s_node
	// Label: nodepool / Value (name of nodepool)
	// Label: zone / Value (name of zone)
	// Label: type / Values: (set of conditions)
	// Label: status / Values: (True/False/Unknown)
	// Value: 1
	RecordNodeConditions(ctx context.Context, labels []map[string]string)

	// node/available
	// Entity: k8s_node
	// Label: nodepool / Value (name of nodepool)
	// Label: zone / Value (name of zone)
	// Label: available / Value (True/False)
	// Label: ready / Value (True/False)
	// Label: scheduleable / Value (True/False)
	// Label: done_warming / Value (True/False)
	// Value: 1
	RecordNodeAvailability(ctx context.Context, labels map[string]string)
}

type ProbeRecorder interface {
	RecordDNSLookupLatency(elapsed time.Duration)
	RecordHTTPGetLatency(statusCode int, elapsed time.Duration)

	// cluster/probe/addon
	// Entity: k8s_cluster
	// Label: name / Value: (name of addon)
	// Label: condition / Value: (Healthy/Unhealthy)
	// Label: reason / Value: Err messages
	// Value: 1
	RecordAddonHealth(ctx context.Context, labels []map[string]string)
}

type Provider interface {
	ClusterRecorder() ClusterRecorder
	NodeRecorder() NodeRecorder
	ProbeRecorder() ProbeRecorder
	Close() error
}