DataDog/datadog-lambda-extension v89

on GitHub

Read about the Next Generation Datadog Lambda Extension.

# AWS Commercial
arn:aws:lambda:<AWS_REGION>:464622532012:layer:Datadog-Extension:89
arn:aws:lambda:<AWS_REGION>:464622532012:layer:Datadog-Extension-FIPS:89
arn:aws:lambda:<AWS_REGION>:464622532012:layer:Datadog-Extension-ARM:89
arn:aws:lambda:<AWS_REGION>:464622532012:layer:Datadog-Extension-ARM-FIPS:89

# AWS GovCloud
arn:aws-us-gov:lambda:us-gov-<AWS_REGION>:002406178527:layer:Datadog-Extension:89
arn:aws-us-gov:lambda:us-gov-<AWS_REGION>:002406178527:layer:Datadog-Extension-FIPS:89
arn:aws-us-gov:lambda:us-gov-<AWS_REGION>:002406178527:layer:Datadog-Extension-ARM:89
arn:aws-us-gov:lambda:us-gov-<AWS_REGION>:002406178527:layer:Datadog-Extension-ARM-FIPS:89

What's Changed

Features

• DogStatsD Improvements

  • Added support for DD_STATSD_METRIC_NAMESPACE environment variable to prefix custom metrics with a namespace (#920, #921 by @jchrostek-dd and @litianningdatadog)

• SnapStart Support

  • Added automatic creation of SnapStart spans and metrics when Lambda SnapStart is enabled (#911 by @jchrostek-dd)
  • Replaces init spans with restore spans for SnapStart invocations

• Trace Agent Improvements

  • Increased trace payload limit from 2 MiB to 50 MiB to prevent connection errors with large trace payloads (#902 by @lym953)
  • Implemented libdatadog's native concentrator for trace stats aggregation (#887 by @lym953)
  • Removed hostname from ClientStatsPayload for proper stats aggregation (#905 by @lym953)

• Secrets Management

• SSM Parameter Store: Add support for retrieving Datadog API keys from AWS Systems Manager Parameter Store via the DD_API_KEY_SSM_ARN environment variable. Supports both String and SecureString parameter types. (#924 by @jchrostek-dd)

• Tag Handling

  • Updated DD_TAGS parsing to support both comma (,) and space ( ) delimiters, improving compatibility with container images (#916 by @jchrostek-dd)

Bug Fixes

• Logging

  • Fixed log processing rules incorrectly filtering orphan logs (extension logs and error logs) when exclude patterns were configured (#912 by @lym953)
  • Add URL information to proxy flusher error logs for easier identification of failing endpoints (#933 by @litianningdatadog)

• Metrics

  • Fix an issue where negative CPU utilization values could be reported due to timing differences between /proc/uptime and /proc/stat (#929 by @shreyamalpani and @litianningdatadog)
  • Fix an issue where the tmp_free metric was missing after the enhanced metrics refactor. (#937 by @shreyamalpani)

Performance

• Removed invocation lock and replaced with aggregator service pattern (#894 by @jchrostek-dd )
• Refactored enhanced metrics monitoring to use long-running task instead of spawning new tasks per invocation (#901 by @shreyamalpani)
• Migrated trace aggregation to channel-based approach to avoid locking (#909 by @shreyamalpani)

Error Handling

• Improved trace agent error messages with additional context (#900 by @lym953)

Known Issues

• Trace stats may show up to 9% under/over-counting in high-volume, high-concurrency scenarios (under investigation)

Full Changelog: v88...v89