Ranger Admin Prometheus / Metrics Configuration Properties¶

Overview¶

Ranger Admin exposes metrics for monitoring and alerting. In large environments, frequent scrapes can trigger expensive database queries. These properties allow you to cache most Ranger Admin metrics and refresh them on a schedule, reducing database load.

Note

The metrics cache applies to admin metrics except database pool metrics. DB pool (c3p0) metrics remain available even when the metrics fetcher is disabled.

All properties that control Ranger Admin Prometheus and metrics behavior. Set these in ranger-privacera-site.xml (or equivalent).

Prometheus metrics (global)¶

Property	Type	Default	Description
`ranger.prometheus.metrics.enabled`	boolean	`false`	When `true`, Prometheus metrics are enabled and `/metrics/prometheus` and `/metrics/db` expose metrics. When `false`, those endpoints return “Prometheus metrics are disabled”.
`ranger.prometheus.http.metrics.enabled`	boolean	`true`	When `true` (and Prometheus metrics are enabled), HTTP request metrics (latency/status) are recorded.
`ranger.jvm.metrics.application.name`	string	`ranger`	Application name tag added to JVM/process metrics (for example \(application="ranger"\)).

Metrics cache (`RangerMetricsFetcher`)¶

DB-backed metrics (user/group/role/tag counts, policies, etc.) can be served from a cache that is refreshed in the background. These properties control that cache.

Add/update the following properties in Ranger Admin configuration (ranger-admin-site.xml).

Property	Type	Default	Description
`ranger.metrics.fetcher.enabled`	boolean	`true`	When `true`, the metrics fetcher and cache scheduler run. When `false`, the cache scheduler is not started and all metric getters return empty or zero (except C3P0/database pool metrics).
`ranger.metrics.cache.enabled`	boolean	`true`	When `true`, admin metrics (except DB pool metrics) are refreshed on a schedule and returned from cache, avoiding expensive DB queries on every scrape. When `false`, every scrape runs live DB queries.
`ranger.metrics.cache.interval.seconds`	long	`3600`	Interval in seconds between cache refreshes. Ignored if `ranger.metrics.cache.enabled` is `false`.
`ranger.metrics.cache.initial.delay.seconds`	long	`0`	Delay in seconds before the first metrics cache refresh. Subsequent refreshes run every `ranger.metrics.cache.interval.seconds`. Ignored if cache or fetcher is disabled. Set to `0` to populate cache soon after startup.
`ranger.metrics.cache.intx.sleep.ms`	long	`300000`	Sleep in milliseconds after each DB read during a cache refresh (to reduce DB load). Set to `0` to disable; useful if you want the first snapshot to complete quickly after restart.

Property	Type	Default	Description
`ranger.metrics.active.plugin.reset.interval`	long	`10`	Interval in minutes for the scheduled task that resets active-plugin metrics.

Summary table (quick reference)¶

Property	Default	Purpose
`ranger.prometheus.metrics.enabled`	`true`	Enable Prometheus metrics endpoints
`ranger.prometheus.http.metrics.enabled`	`true`	Enable HTTP request metrics
`ranger.jvm.metrics.application.name`	`ranger`	Tag for JVM/process metrics
`ranger.metrics.fetcher.enabled`	`true`	Enable metrics fetcher and cache scheduler
`ranger.metrics.cache.enabled`	`true`	Use cache for DB-backed metrics
`ranger.metrics.cache.interval.seconds`	`3600`	Interval between cache refreshes (seconds)
`ranger.metrics.cache.initial.delay.seconds`	`0`	Delay before first cache refresh (seconds)
`ranger.metrics.cache.intx.sleep.ms`	`300000`	Sleep after each DB read during refresh (ms); use `0` for faster first snapshot
`ranger.metrics.active.plugin.reset.interval`	`10`	Active-plugin metrics reset interval (minutes)

Example (`ranger-admin-site.xml`)¶

XML
<property>
  <name>ranger.metrics.cache.enabled</name>
  <value>true</value>
  <description>When true, admin metrics (except DB pool metrics) are refreshed on a schedule and returned from cache; avoids expensive DB queries on every scrape.</description>
</property>
<property>
  <name>ranger.metrics.cache.interval.seconds</name>
  <value>3600</value>
  <description>Interval in seconds for background refresh of metrics cache. Ignored if ranger.metrics.cache.enabled is false.</description>
</property>
<property>
  <name>ranger.metrics.cache.initial.delay.seconds</name>
  <value>0</value>
  <description>Delay in seconds before the first metrics cache refresh. Subsequent refreshes run every ranger.metrics.cache.interval.seconds. Ignored if cache or fetcher is disabled.</description>
</property>
<property>
  <name>ranger.metrics.cache.intx.sleep.ms</name>
  <value>300000</value>
  <description>Sleep in milliseconds after each metrics read during cache refresh, to yield to other DB activity. Set to 0 to disable.</description>
</property>
<property>
  <name>ranger.metrics.fetcher.enabled</name>
  <value>true</value>
  <description>When false, RangerMetricsFetcher is disabled: the metrics cache scheduler is not started and all metric getters return empty or zero. C3p0/database pool metrics (getDatabaseMetrics) remain available.</description>
</property>

Initial delay / first scrape behavior¶

When the metrics cache is enabled, some metrics can be empty/zero immediately after Ranger Admin starts until the first background refresh completes (and depending on ranger.metrics.cache.initial.delay.seconds, ranger.metrics.cache.interval.seconds, and ranger.metrics.cache.intx.sleep.ms).

Prev topic: Advanced Configurations