Ingest additional metrics

Scraping metric endpoints is the preferred method of ingesting metrics to the Collector, but you can also ingest the following metric formats:

• DogStatsD
• StatsD
• Carbon
• Prometheus and OpenMetrics
• Pushgateway
• SignalFx
• Wavefront

Enable aggregations

The v0.101.0 release adds an option to enable client-side aggregations for Carbon, StatsD, and Datadog metrics. These aggregations reduce network egress and processed writes to the Chronosphere app.

When the Collector ingests a data sample, it aggregates the data based on the metric type and sends a single data point to the Chronosphere app that represents the sample for the defined period. The Collector aggregates on the following metric types:

• Counters, which are rolled up through addition.
• Gauges, which are selected by the last value written.
• Timers, which use sketches of the underlying sample to build quantiles.

Aggregations are disabled by default. To enable aggregations, define the push.PROTOCOL.aggregation YAML collection in your Collector manifest, where PROTOCOL can be carbon, dogstatsd, or statsd. For example, the following YAML collection enables aggregations for StatsD with defined values for each of the supported metric types:

    push:
       statsd:
         aggregation:
           counters:
             interval: "10s"
           gauges:
             interval: "10s"
           timers:
             interval: "10s"
           inactiveExpireAt: "2m"

• interval: Defines the length of time before writing to the Chronosphere app. A longer interval duration provides more effective caching, but also increases the length of delays before writing.

• inactiveExpireAt: Determines the length of time before Chronosphere evicts unused keys from the local cache. Default: 2m.