Request for sending data via telemetry

The goal is to collect metrics about Cluster Logging Operator 6.y, so that we can track usage of features in the new release version.

openshift_logging:log_forwarder_pipelines:sum

"openshift_logging:log_forwarder_pipelines:sum" represents the number of logging pipelines managed by CLO per namespace.

Labels

• "resource_namespace", the namespace in which the forwarder is deployed. Most customers only use "openshift-logging".
• "version", the version of the Cluster Logging Operator

The cardinality of the metric is "one per namespace", which for most clusters will be one.

openshift_logging:log_forwarder_pipelines:count

"openshift_logging:log_forwarder_pipelines:count" represents the number of deployed ClusterLogForwarders per namespace.

Labels

• "resource_namespace", the namespace in which the forwarder is deployed. Most customers only use "openshift-logging".
• "version", the version of the Cluster Logging Operator

The cardinality of the metric is "one per namespace", which for most clusters will be one.

openshift_logging:log_forwarder_input_type:sum

"openshift_logging:log_forwarder_input_type:sum" represents the number of inputs managed by CLO per namespace.

Labels

• "resource_namespace", the namespace in which the forwarder is deployed. Most customers only use "openshift-logging".
• "version", the version of the Cluster Logging Operator
• "input", the type of input used. There are four input types.

The cardinality of the metric is "one per namespace and input type". I expect this to be two for most customers.

openshift_logging:log_forwarder_output_type:sum

"openshift_logging:log_forwarder_output_type:sum" represents the number of outputs managed by CLO per namespace.

Labels

• "resource_namespace", the namespace in which the forwarder is deployed. Most customers only use "openshift-logging".
• "version", the version of the Cluster Logging Operator
• "output", the type of output used. There are eleven output types.

The cardinality of the metric is "one per namespace and output type". I expect most customers to use one or two output types.

openshift_logging:vector_component_received_bytes_total:rate5m

"openshift_logging:vector_component_received_bytes_total:rate5m" represents current total log rate for a cluster for log collectors managed by CLO.

Labels

• "namespace", the namespace in which the forwarder is deployed. Most customers only use "openshift-logging".

The cardinality of the metric is "one per namespace". which for most clusters will be one.

Links

Component exposing the metric: https://github.com/openshift/cluster-logging-operator/blob/master/internal/metrics/telemetry/telemetry.go#L25-L47

The recording rules for these metrics are currently reviewed in this PR: https://github.com/openshift/cluster-logging-operator/pull/2823