Histogram

Histograms visualize data distribution in different range groups. Sysdig Monitor supports three types of Histograms.

  1. The Histogram panel type on the Dashboard: Histogram panels allow you to visualize the distribution of metric values for large datasets. The histogram panel helps understand value across different segments. For example, CPU usage percent by containers across your hosts gives you the aggregated value across the selected time. Use Histograms for any metric, Sysdig native or custom, counter or gauge, segmented by a dimension/label.

    To create a Histogram panel, Create a new dashboard or select an existing dashboard and click Add Panel. Select the visualization icon, which defaults to Timechart. Under Select Visualization, choose Histogram. Customize your Histogram with the Metric & Grouping controls, ensuring you select at least one label. When you’re done, click Save.

  1. Legacy Prometheus histogram collection: This implementation of legacy Prometheus Histograms is deprecated in SaaS 3.2.6 release.

    To create a Histogram, use the Prometheus integration to collect histogram metrics and use the PromQL panel with the histogram_quantile function.

    Histograms are not reported by default in Prometheus. This is configured in dragent.default/yaml:

prometheus:
  # enabled: false
  interval: 10
  log_errors: true
  max_metrics: 1000
  max_metrics_per_process: 1000
  max_tags_per_metric: 40
  histograms: false
  timeout: 1

To report histogram metrics, edit this to read histograms: true.

  1. Prometheus histograms (collected as raw metrics): The legacy Prometheus histogram collection is replaced by the new Prometheus histogram. You can natively collect histogram metrics, and for visualization, use timechart:

    For example, run the following query to build a timechart:

    sum(histogram_metrics_bucket{kubernetes_cluster_name="prod"}) by (le)