Connectware Metrics

How Connectware exposes metrics, which components are covered, and how the metrics integrate with external monitoring tools.

Connectware exposes metrics in OpenMetrics format from its key components, so you can monitor connection health, data flow, service state, and authentication activity with the tools you already use, such as Prometheus, Grafana, or Datadog.

Each component serves a /metrics endpoint that, for example, a Prometheus server scrapes at a configured interval and stores in its time series database for dashboards and alerting. Metrics carry labels that break down measurements further, for example a connection count by protocol and status, or a message counter by operation type.

Monitored Components

The following components expose metrics.

Component	What the metrics tell you
Protocol Mapper	Whether connections to machines are established, how much data is flowing through endpoints and mappings, and whether node operations are executing correctly.
Service Manager	How many services are running or disabled, whether recent installations succeeded, how long service enabling takes, and whether dependency resolution failures are occurring.
Resource Status Tracking	Whether any services are in a deviated state — that is, not matching their expected configuration.
Auth Server	Whether authentication and authorization requests are succeeding, and where failures are occurring by type.

For the full list of individual metrics, their types, labels, and example output for each component, see Metrics Reference.

Broker and NATS expose metrics endpoints as well, but their metrics come from the underlying software rather than from Connectware instrumentation. For Broker metrics, see Monitoring & Analytics. NATS exposes its metrics through a Prometheus exporter sidecar that is deployed by default and translates NATS metrics into an OpenMetrics-compatible format. Set nats.metrics.enabled to false to disable it.

Prometheus Operator Integration

On Kubernetes, the connectware and connectware-agent Helm charts can create Prometheus Operator ServiceMonitor resources for each component, removing the need to maintain a static scrape configuration. See Configuring Metrics for the Helm configuration reference.

Rate Limiting and Collection Interval

Each metrics endpoint has a built-in rate limiter. When the limit is reached, the endpoint returns the last cached value rather than rejecting the request. Metric values are updated on a fixed collection interval — scraping more frequently than this interval returns the same cached values. The default collection interval is 5,000 milliseconds.