Deployment Integrations
On-premise Integrations
Platform Integrations

Metrics Configurations

Provided Metrics

Metric support provides all essential information regarding the performance of your Java Lambda functions, which translate to carefully calculated stats. It lets you examine:

  • Loaded class counts

  • Memory usages

  • Memory usages by pool

  • CPU percentages

  • GC CPU percentages

  • GC counts

  • GC durations

  • Thread count

Configuring Metrics

If you use Thundra Java layer (or custom runtime), metric plugin is disabled by default but you can enable it by setting thundra_agent_lambda_metric_disable environment variable to false.

Metrics are periodically provided and reported. With the following environment variables, the reporting period can be configured. If one of the (no need to all of them) following periods reach its configured limit, the metric is published:

  • thundra_agent_metric_sample_sampler_timeAware_timeFreq: Configures the time-frequency in milliseconds to to trigger metric reporting. Metrics are provided only once during the specified time period. The default value is 300000 milliseconds (5 minutes).

  • thundra_agent_metric_sample_sampler_countAware_countFreq: Configures the invocation count to trigger metric reporting. Metrics are provided only once at each of the specified invocation count. The default value is 100.