The OpenTelemetry Collector Contrib contains everything in the opentelemetry-collector release, be sure to check the release notes there as well.

Unmaintained Components

These components are marked as Unmaintained and will eventually be removed from our community distributions. If you depend on these components we need your help to support them.

• exporter/alibabacloudlogservice
• receiver/carbon

End User Changelog

🛑 Breaking changes 🛑

• exporter/elasticsearch: Remove ecs mode span enrichment for span.action, span.message.queue.name, and transaction.message.queue.name (#45014) The removed span enrichments have been moved to the github.com/elastic/opentelemetry-collector-components/processor/elasticapmprocessor. It is recommended to use the github.com/elastic/opentelemetry-collector-components/processor/elasticapmprocessor when using mapping mode ecs to ensure index documents contain all required Elastic fields to power the Kibana UI.

• exporter/kafka: Remove Sarama producer implementation (#44565) The Sarama-based Kafka producer has been removed from kafkaexporter. Feature gate exporter.kafkaexporter.UseFranzGo has also been removed since Franz-go is now the only supported Kafka client.

• processor/tail_sampling: The deprecated invert decisions are disabled by default. (#44132) Drop policies should be used instead of invert decisions for explicitly not sampling a trace. If the deprecated behavior is required while migrating to drop policies, disable the processor.tailsamplingprocessor.disableinvertdecisions feature gate.

• receiver/kafka: Remove Sarama consumer implementation and default_fetch_size configuration option (#44564) The Sarama-based Kafka consumer has been removed from kafkareceiver. The default_fetch_size configuration option has also been removed as it was only used by the Sarama consumer. Feature gate receiver.kafkareceiver.UseFranzGo has also been removed since Franz-go is now the only supported Kafka client.

🚩 Deprecations 🚩

• exporter/elasticsearch: Deprecate mapping::mode config option (#45246) The mapping::mode config option is now deprecated and will soon be ignored. Instead, use the X-Elastic-Mapping-Mode client metadata key (via headers_setter extension) or the elastic.mapping.mode scope attribute to control the mapping mode per-request. See the README for migration instructions.

🚀 New components 🚀

• processor/lookup: Add skeleton for external lookup enrichment processor (#41816) Adds the initial skeleton for a lookup processor that performs external lookups to enrich telemetry signals. Also includes source abstraction with factory pattern, noop source for testing, and cache wrapper utility.

💡 Enhancements 💡

• cmd/schemagen: Extend schemagen script with ability to handle external refs. (#42214) The schemagen tool has been enhanced to support external references when generating configuration schemas for OpenTelemetry Collector components. This improvement allows the tool to accurately reference and include schema definitions from external packages, facilitating better modularity and reuse of configuration schemas across different components.

• cmd/schemagen: Fixes for schemagen to handle common issues with receiver components schemas. (#42214) Fix common issues discovered while using schemagen with receiver components:

  • Missing config.go file (e.g. namedpipereceiver)
  • Parsing obsolete types (e.g. nsxtreceiver)
  • Unable to embed fields with squash tag and not exported internal type (e.g. huaweicloudcesreceiver)

• cmd/telemetrygen: Add batching capability to metrics and traces (#42322)

  • Changed traces batching to have configurable batch size and match batch flag.
  • Added batching to metrics.
  • Added batching to logs.

• exporter/azureblob: Add timezone option for formatting blob names in azureblob exporter. (#43752)

• exporter/elasticsearch: Remove go-elasticsearch dependency to reduce binary size (#45104) This leads to a 19MB size reduction in contrib distribution

• exporter/googlecloudstorage: Add support for time partitioning (#44889)

• exporter/opensearch: Add support for multiple variables to build index names (#42585)

• exporter/sumologic: Add decompose_otlp_summaries configuration option to decompose OTLP Summary metrics into individual gauges and counters (#44737)

• extension/awslogs_encoding: Optimize CloudTrail logs unmarshaling for memory usage (#45180)

• processor/k8sattributes: Bumnp version of semconv to v1.39.0 (#45447)

• processor/redaction: Add sanitize_span_name option to URL and DB sanitization configs. (#44228)

• processor/redaction: Add ignored_key_patterns configuration option to allow ignoring keys by regex pattern (#44657)

• processor/resourcedetection: Add optional docker attributes (#44898) Add container.image.name and container.name optional resource attributes with the docker detector. . Note: Because of opentelemetry-collector-releases#1350 this change is not available on the v0.144.0 binary releases. v0.145.0 will include this change.

• processor/tail_sampling: Provide an option, decision_wait_after_root_received, to make quicker decisions after a root span is received. (#43876)

• receiver/azureeventhub: Add support for azure auth when feature gate receiver.azureeventhubreceiver.UseAzeventhubs is enabled. (#40711)

• receiver/prometheus: receiver/prometheus now associates scraped _created text lines as the created timestamp of its metric family rather than its own metric series, as defined by the OpenMetricsText spec (#45291)

• receiver/prometheus: Add comprehensive troubleshooting and best practices guide to Prometheus receiver README (#44925) The guide includes common issues and solutions, performance optimization strategies, production deployment best practices, monitoring recommendations, and debugging tips.

• receiver/prometheusremotewrite: Replace labels.Map() iteration with direct label traversal to eliminate intermediate map allocations. (#45166)

🧰 Bug fixes 🧰

• exporter/kafka: franz-go: Exclude non-produce metrics from kafka_exporter_write_latency and kafka_exporter_latency (#45258)

• exporter/opensearch: Fix dynamic log index feature putting logs in wrong index (#43183)

• exporter/prometheusremotewrite: Prevent duplicate samples by allowing the WAL to be empty (#41785) Since the WAL is being truncated after every send it's likely the reader and writer are in sync. Since WAL was not allowed to be empty, the reader would always re-read previously delivered samples causing duplicate data to be sent continuously.

• extension/datadog: Datadog extension no longer throws an error for missing extensions when getting a list of active components, and now populates active components even when missing go mod/version info. (#45358, #45460)

• extension/file_storage: Handle filename too long error in file storage extension by using the sha256 of the attempted filename instead. (#44039)

• extension/text_encoding: Avoid spurious marshalling separators at end of lines (#42797) Previously, text_encoding would append the marshalling separator to the end of each log record, potentially resulting in double-newlines between blocks of records.

• extension/text_encoding: Fix an issue where marshalling/unmarshalling separators were ignored (#42797)

• pkg/kafka/configkafka: Fix consumer group rebalance strategy validation (#45268)

• pkg/ottl: Fix numeric parsing to correctly handle signed numbers in math expressions. (#45222) The OTTL math expression parser did not correctly handle unary signs for plus and minus. Expressions like 3-5 would not parse correctly without inserting spaces to make it 3 - 5. This change moves the sign handling out of the lexer and into the parser.

• pkg/ottl: Handle floating constants with decimal point but no fraction. (#45222) Floating point constants that had a decimal point but no fractional digits (e.g., "3.") were not handled properly and could crash the parser. These are now parsed as valid floating point numbers.

• pkg/stanza: Fix Windows UNC network path handling in filelog receiver (#44401) The filelog receiver now correctly handles Windows UNC network paths (e.g., \server\share\logs*.log). Previously, the receiver could list files from network shares but failed to open them due to path corruption during normalization. This fix converts UNC paths to Windows extended-length format (\?\UNC\server\share\path) which is more reliable and not affected by filepath.Clean() issues.

• pkg/stanza: Ensure container parser respects the if condition and on_error settings when format detection fails (#41508)

• processor/resourcedetection: Prevent the resource detection processor from panicking when detectors return a zero-valued pdata resource. (#41934). Note: Because of opentelemetry-collector-releases#1350 this change is not available on the v0.144.0 binary releases. v0.145.0 will include this change.

• processor/resourcedetection: Fix nil pointer panic when HTTP client creation fails in Start method (#45220). Note: Because of opentelemetry-collector-releases#1350 this change is not available on the v0.144.0 binary releases. v0.145.0 will include this change.

• receiver/awslambda: Fix S3 key usage in AWS Lambda Receiver (#45364)

• receiver/datadog: Fix service check endpoint to handle both array and single object payloads (#44986) The /api/v1/check_run endpoint now uses defensive parsing to handle both array [{...}] and single object {...} payloads. This fixes intermittent unmarshal errors when Datadog agent sends connectivity health checks.

• receiver/jmx: Enable initial_delay and collection_interval settings via scraper helper (#44492)

• receiver/libhoney: Improve msgpack decoding to handle ints or uints (#45273)

• receiver/postgresql: Fix query plan EXPLAIN to use raw query with $N placeholders instead of obfuscated query with ? placeholders (#45190) Previously, the EXPLAIN query was using obfuscated queries with ? placeholders, which PostgreSQL does not recognize. Now uses the raw query with $1, $2 placeholders that PostgreSQL expects.

• receiver/prometheusremotewrite: Fix silent data loss when consumer fails by returning appropriate HTTP error codes instead of 204 No Content. (#45151) The receiver was sending HTTP 204 No Content before calling ConsumeMetrics(), causing clients to believe data was successfully delivered even when the consumer failed. Now returns 400 Bad Request for permanent errors and 500 Internal Server Error for retryable errors, as per the Prometheus Remote Write 2.0 specification.

• receiver/sqlserver: Accuracy improvements for top-query metrics (#45228) SQLServer metrics reporting is improved by reducing the warm-up delay and providing accurate insights sooner.

API Changelog

🛑 Breaking changes 🛑

• all: add unix socket support to HTTP server components (#45308) HTTP server components (namely receivers) now support listening on Unix domain sockets in addition to TCP addresses, by configuring transport: unix and setting endpoint to a Unix socket path.

  This is a breaking change to the config structs, but is not breaking for end users. Existing YAML configuration options remain unchanged.

• receiver/aerospike: Move mock structs to internal package. (#43995)

We are thrilled to welcome our first-time contributors to this project. Thank you for your contributions @jkoronaAtCisco, @gustavo89587, @zoelandeyro, @danielbcorreia, @mbdeveci, @arminsumo, @eriksywu, @dmytrohysht, @jon-signal, @cruscio, @Confushion ! 🎉