KNOWLEDGE BASE

Release Notes v5.0.0

Fluent Bit is a Fast and Lightweight Telemetry Agent for Linux, BSD, macOS and Windows. We are proud to announce the availability of Fluent Bit v5.0.0.

For people upgrading from previous versions you must read the Upgrading Notes section of our documentation:

https://docs.fluentbit.io/manual/5.0/installation/upgrade_notes

Introduction

Fluent Bit, a CNCF graduated project under the umbrella of Fluentd, announces the availability of v5.0.0.

Fluent Bit v5.0 is a major release focused on scale, correctness, and modern telemetry integration. Compared to the previous stable line, this release improves HTTP ingestion concurrency, expands OpenTelemetry and OAuth2 coverage, adds new processing and compression capabilities, and strengthens the operational signals that production teams rely on.

Between v4.2.3 and v5.0.0, the project landed more than 400 commits covering performance work, protocol compatibility, security improvements, and hardening across inputs, outputs, processors, and internal libraries.

🔥 Highlights

Faster and More Scalable HTTP Ingestion

The biggest operational theme in v5.0 is better concurrency for HTTP-based ingestion. Worker runtime support was added across the shared HTTP server and adopted by multiple receivers, including:

• http
• splunk
• elasticsearch
• opentelemetry
• prometheus_remote_write

This work improves throughput on busy collectors, reduces contention in high-volume gateways, and helps Fluent Bit scale much more efficiently under OTLP-heavy workloads. When receiving OTLP payloads, Fluent Bit v5.0 is up to 4x faster than Fluent Bit 4.2.

OpenTelemetry Coverage Across Logs, Metrics, and Profiles

OpenTelemetry received broad investment in this release:

• broader compatibility across logs, metrics, and profiles
• better handling of evolving OpenTelemetry payload formats
• stronger metrics support, including exponential histograms
• smoother interoperability with modern OpenTelemetry deployments

The result is a more capable and more future-ready OpenTelemetry path for teams standardizing on OTLP across their telemetry pipelines.

OAuth2 and Secure Integration Improvements

Fluent Bit v5.0 also expands secure integration options across the internal stack, extending OAuth2 support through shared components, HTTP clients, inputs, and outputs:

• config-map based OAuth2 support for outputs
• OAuth2 client credentials and private_key_jwt support in out_http
• OAuth2 JWT validation in in_http
• improved JWKS handling, cache coordination, retry behavior, and cleanup paths
• stronger teardown and failure handling in OAuth2-protected output flows

These changes make Fluent Bit easier to deploy in zero-trust and enterprise identity environments without requiring sidecars or external auth shims.

More Accurate Operational Accounting

Several changes in v5.0 continue the work to make Fluent Bit’s operational signals more representative and more useful at scale:

• better logical record accounting in outputs such as S3
• clearer metrics around grouped logs, retries, and routed traffic
• improved worker ownership and ingress flow tracking
• new backpressure duration metrics for earlier visibility into sustained pressure
• a new JSON health endpoint in HTTP server API v2

This work matters for teams that use Fluent Bit metrics for SLOs, chargeback, delivery validation, or troubleshooting under load.

New Capabilities

• processor_cumulative_to_delta: a new metrics processor that converts cumulative metrics to delta values, including exponential histogram support.
• processor_tda: a new topological data analysis processor based on Ripser for advanced metrics analysis workflows.
• Internal logs input: a new in_fluentbit_logs plugin to route Fluent Bit’s own logs through standard pipelines.
• Compression options: Snappy support was added across shared AWS compression paths, out_s3, and out_kinesis_firehose; compression support was added to out_kinesis_streams; and out_azure_blob gained Zstandard support.
• Windows event rendering: in_winevtlog now supports text rendering for collected events.
• Tail observability: in_tail now exposes a skipped-lines counter and includes database-operation hardening for busy environments.

Important Fixes and Hardening

• HTTP-based inputs now preserve busy and ingest-error behavior more consistently under worker load.
• in_http, in_splunk, and in_elasticsearch received fixes for payload/error handling and response status preservation.
• in_emitter limits filesystem growth under downstream backpressure.
• in_storage_backlog preserves DLQ data across restarts.
• Kubernetes metadata extraction was hardened to prevent expensive regex backtracking.
• MQTT, TCP, HTTP, and OAuth2 paths received multiple cleanup and exception-safety fixes that reduce crash, leak, and invalid-state risk.

Performance and Platform Work

• Shared HTTP server worker support improves effective concurrency for ingestion-heavy deployments.
• OpenTelemetry input parsing reduces temporary-buffer overhead in some hot paths.
• SIMD-based packing improvements continue to improve throughput on supported architectures.
• Backpressure duration metrics provide better observability into sustained pressure before it becomes a delivery problem.

Libraries

• cmetrics: upgrade through the v2.0.x line, including exponential histogram support
• cprofiles: upgrade to v0.2.1
• ctraces: upgrade to v0.7.0
• fluent-otel-proto: upgrade to v0.12
• monkey: upgrade to v1.8.7

Join us

We want to hear from you, our community keeps growing and you can be part of it:

• Github: http://github.com/fluent/fluent-bit
• Slack: http://slack.fluentd.org
• Twitter: @fluentbit