HashiConf Last chance to register for our cloud conference October 10-12. Register today
Nomad
Nomad Home

dynamic_application_sizing Block

Placementdynamic_application_sizing
Enterprise

This functionality only exists in Nomad Autoscaler Enterprise. This is not present in the open source version of Nomad Autoscaler.

Note that currently Prometheus is the only APM availible for Dynamic Application Sizing

The dynamic_application_sizing blocks configures the Nomad Autoscaler agent options for Dynamic Application Sizing. Learn more about the Dynamic Application Sizing functionality in the Dynamic Application Sizing Concepts guide

dynamic_application_sizing {
  evaluate_after            = "1h"
  metrics_preload_threshold = "12h"
}

dynamic_application_sizing Parameters

  • cpu_metric (string: "nomad_client_allocs_cpu_total_ticks") - The APM metric to use for querying the historical CPU usage of a task.

  • evaluate_after (string: "24h") - Specifies how much historical data must be available for a task before the Autoscaler starts generating recommendations.

  • group_label (string: "task_group") - The label used by the APM to store the name of a task group. It is used by the Autoscaler when querying the APM for metrics.

  • job_label (string: "exported_job") - The label used by the APM to store the ID of a job. It is used by the Autoscaler when querying the APM for metrics.

  • memory_metric (string: "nomad_client_allocs_memory_usage") - The APM metric to use for querying the historical memory usage of a task.

  • metrics_preload_threshold (string: "168h") - Specifies how much historical data the Autoscaler will preload from the APM when it starts.

  • namespace_label (string: "namespace") - The label used by the APM to store the namespace of a job. It is used by the Autoscaler when querying the APM for metrics.

  • task_label (string: "task") - The label used by the APM to store the name of a task. It is used by the Autoscaler when querying the APM for metrics.