KNOWLEDGE BASE

Release Notes v5.0.7

Fluent Bit is a Fast and Lightweight Data Processor and Forwarder for Linux, BSD, Windows and macOS. We are proud to announce the availability of Fluent Bit v5.0.7.

This maintenance release improves parser time-zone handling, strengthens HTTP server worker lifecycle behavior, adds new eBPF, S3, Chronicle, OAuth2, and Windows input capabilities, and hardens protocol, compression, storage, and format handling across production pipelines.

What’s new ?

Core

• parser: add time_zone support for IANA time zones on timestamps without embedded time-zone data, including Windows/IANA mapping, local zoneinfo validation, missing-year handling, and cached lookups
• http_server: improve multi-worker startup and teardown by serializing runtime cleanup, keeping worker exit callbacks on the worker thread, protecting downstream registration, and avoiding connection-drop races
• http_client: format HTTP/2 authority headers correctly for IPv6 literal hosts
• storage: discard checksum-corrupt chunks when irrecoverable chunk deletion is enabled
• compression: improve Snappy framed-data validation and add bounded Zstd decompression handling
• avro: preserve integer range errors and handle 64-bit integer values correctly when converting MessagePack to Avro
• config_format: improve YAML variant cleanup on parse errors
• network: treat SO_REUSEPORT as a no-op on Windows, improving cross-platform listener setup
• ci: pin GitHub Actions workflows to SHA-backed actions and update lint review automation
• tests: expand internal, runtime, and integration coverage for parser time zones, HTTP server workers, HTTP/2 IPv6, storage checksum cleanup, OAuth2 user-agent headers, S3 quarantine behavior, Kinesis event limits, Chronicle metadata mapping, Avro integers, and protocol validation

Libraries

• chunkio: upgrade the bundled library to v1.5.5 with improved handling for checksum-corrupt chunks during recovery scans

Plugins

• eBPF (Input)
  • add a dns trace for DNS query and response observability
• STDIN (Input)
  • enable the plugin on Windows and add Windows-aware standard input handling for pipes and console input
• Tail (Input)
  • fix offset_key values on Windows 64-bit builds
  • add integration coverage for parser time_zone handling and checksum-corrupt filesystem storage chunks
• Storage Backlog (Input)
  • remove checksum-corrupt chunks when storage.delete_irrecoverable_chunks is enabled
• S3 (Output)
  • add retry_exhausted_action so chunks that exceed retry_limit can be quarantined or deleted
  • add quarantine_dir_limit_size to cap disk usage for retry-exhausted quarantined chunks
• Chronicle (Output)
  • add namespace, namespace_key, and repeated label settings for native out-of-band namespace and label mapping
  • split batches when resolved namespace or label metadata changes, preserving Chronicle payload consistency
• HTTP (Output)
  • add oauth2.user_agent for setting a User-Agent header on OAuth2 token requests
• Kinesis Streams (Output)
  • increase the PutRecords payload limit to 10 MB and add event-size boundary coverage
• OpenTelemetry (Output)
  • improve HTTP/2 delivery to IPv6 literal endpoints
• Stackdriver (Output)
  • improve metadata and request cleanup on failure paths
• Syslog (Input)
  • harden octet-counted frame length handling for malformed input
• Forward (Input)
  • improve validation for message-mode payloads before accessing record fields
• Collectd (Input)
  • improve validation of malformed network protocol parts
• MQTT (Input)
  • improve handling of truncated publish packets
• Elasticsearch (Input)
  • improve bulk request validation for empty metadata maps

Join us

We want to hear about you, our community is growing, and you can be part of it!, you can contact us at:

• Github: http://github.com/fluent/fluent-bit
• Slack: CNCF Slack, channel #fluent-bit (https://communityinviter.com/apps/cloud-native/cncf)
• Twitter: @fluentbit