Metrics

Ondat believes in exposing many metrics about the functioning and performance of Ondat processes to help users instrument applications consuming Ondat volumes and Ondat itself. To this end Ondat exposes metrics via a Prometheus endpoint on each Ondat pod. See our Prometheus Endpoint reference page for specific details about what metrics are exposed.

Ondat metrics are exposed in Prometheus text format, so collectors such as Prometheus, Telegraf or Sensu can be used. Prometheus text format exposes data as time series where each time series can be one of four Prometheus metric types.

Metric Name	Description	Example Metric
Counter	Cumulative metric that only increases. Can be reset to zero on a restart	storageos_volume_backend_read_bytes_total
Gauge	Metric that can increase or decrease	storageos_volume_size_bytes
Histogram	Cumulative metric that includes information about the distribution of samples	storageos_local_leader_known_nodes_sync_seconds
Summary	Similar to a histogram but calculates quantiles over certain time windows	go_gc_duration_second

Each time series is identified by a metric name and key-value labels. The key-value labels allow multiple dimensions for a metric to be exposed. For example the storageos_volume_backend_read_bytes_total metric is given for each volume.

Using Prometheus to scrape metrics endpoints in Kubernetes is quite elegant as Prometheus can be configured to scrape metrics from Kubernetes Services. This is important because Ondat metrics are intended to be scraped from every Ondat pod in the cluster and then aggregated.

For an example of how to visualize Ondat metrics please see our Monitoring Ondat page for a link to a sample Grafana dashboard.