Agent v1.30.0

Metric Aggregation, Extract Metric Node, Datadog Source Node, and Prometheus Source Node.

  2 minute read  

February 24, 2025

New Features

  • Metric Aggregation: Introduced the aggregate metric node to batch and aggregate metrics.
  • Extract Metric Node: Introduced the extract metric node, improving data extraction capabilities in pipelines.
  • Datadog Source Node: Released for more comprehensive data ingestion from Datadog agents.
  • Prometheus Source Node: moved to released status.
  • Sampling Node: moved to released status.
  • SCC Support in Helm: Added support for pod security context configurations in Helm to simplify deployment in OpenShift environments.
  • DNS Overrides: Added support for DNS service configuration environment variable overrides.

Enhancements

  • Self Telemetry Optimization: Optimized self telemetry to reduce CPU and memory consumption. This includes improvements in how telemetry maps are handled and passed.
  • Logging Level Control: Updated self-telemetry settings to use internal logging levels to prevent verbose and unnecessary logs when pipeline log level is set to “debug”.
  • Template Reuse in Logs: Improved support for Demo node format templates to allow reuse of variables within logs and added enhancements for timestamp and number formatting.
  • Node Status and Environment Compatibility: Defined and added support for specific environments in source node configurations, enhancing pipeline configuration validation and operational restrictions.
  • Pattern Clustering with Log Levels: Updated the pattern clustering algorithm to incorporate log levels, enabling more accurate clustering of logs with similar severity.
  • Self-Telemetry Integration: Integrated telemetry reporting across Prometheus input and tailers for enhanced monitoring of inbound and outbound data flows.
  • Kubernetes Metrics Collection: Refactored the include/exclude logic of k8s metrics source node to support regex patterns and improve component-specific metric filtering. The exclude block takes precedence, and if both blocks are empty, only kube_state_metrics will be scraped.

Bug fixes

  • Fix for OTLP Collector Issue: Resolved metric conversion issues between OTLP Collector and OTLP source, ensuring consistent metric handling.
  • Pattern Source Fix: Corrected an issue where the source type was not properly set for generated patterns.
  • Configuration Adjustments: Streamlined configurations within Helm charts to use a unified repository field for container management, reducing redundancy.
  • Metric Discrepancy Resolution: Addressed metric discrepancies by fixing type conversions, ensuring alignment between different metric input streams.
  • OTTL in Route Node: Changed Route node to use standard OTTL syntax when in OTTL mode.