Agent v1.25.0

Amazon CloudWatch Destination Node, Azure Log Analytics Destination Node, Microsoft Sentinel Destination Node, and EDXEncode OTTL Function.

December 17, 2024

New Features

  • Amazon CloudWatch Destination Node: Introduced a new Amazon CloudWatch destination node. This feature includes dynamic namespace, group, and stream name configurations, enhancing integration flexibility.
  • Azure Log Analytics Destination Node: Implemented a new Azure Log Analytics destination node.
  • Microsoft Sentinel Destination Node: Added a new destination node for Azure Sentinel.
  • EDXEncode OTTL Function: Introduced the EDXEncode OTTL function for string-to-byte array conversions, such as modifying log bodies.

Enhancements

  • Service Name Dimension in Metrics: Enhanced the IOMetricCollector to support the service.name dimension in metrics.
  • Generalized NaN and Inf Handling: Extended the logic for discarding NaN and Inf values across all metric-related tailers.
  • Node Health Metrics Update: Added ed.pipeline.node.type to node_health metrics and updated rollup rules to incorporate this new label.
  • Dynamic Resource Type for GCL: Supported dynamic resource type definition for the GCL destination node.
  • Logging Directory Customization: Introduced ED_LOGGING_DIR environment variable to allow users to override the default logging directory of the agent.

Bug Fixes

  • Prometheus Collector Synchronization: Resolved a stop procedure issue in the Prometheus collector by removing unnecessary synchronization variables.
  • Proper Message Splitting for GCL: Improved message splitting for the GCL destination node.

Security

  • CVE-2024-45337 Mitigation: Updated golang/x/crypto package to version 0.31 to address a high-severity CVE.

Miscellaneous

  • Metric Naming and Schema Updates: Renamed metrics and functions to align with OTEL schema changes, including ed.pipeline.<read/write>_lines to ed.pipeline.<read/write>_items, and updated item-related function names for consistency.
  • Internal ED Destination Features Removal: Completed the removal of the features parameter from internal ED outputs, streamlining data handling processes.
  • Logging Optimization: Reduced excessive logging in the Log to Pattern node by implementing error rate limiting and adjusting log levels for operational messages.