Fluentd review 2026: a free, CNCF-graduated unified logging layer with 500+ plugins. See architecture, pricing, pros, cons, and Fluent Bit alternatives.
Fluentd is a free, open-source data collector created to build a unified logging layer for distributed systems. It sits between data sources and backend storage or analytics systems, converting logs into a common JSON format so they can be filtered, buffered, and routed consistently regardless of where they originated.
Originally created in 2011 by Sadayuki Furuhashi at Treasure Data, Fluentd is now a Cloud Native Computing Foundation graduated project, sitting alongside sibling project Fluent Bit under the CNCF Fluent umbrella. It is released under the Apache License 2.0 and has no cost to download, run, or modify.
Fluentd's core strength is its plugin architecture: over 500 input, filter, buffer, and output plugins let it connect to nearly any log source and nearly any destination, from local files and syslog to Kafka, Elasticsearch, S3, and dozens of SaaS platforms.
Its buffering system groups events into memory- or file-based chunks keyed by tag, time, or record fields, with configurable retries and a secondary-output fallback, giving it at-least-once delivery semantics that protect against data loss during backend outages.
Every event that flows through Fluentd carries a tag, and match directives with wildcard support route each tagged event to the correct output configuration, similar to a rules-based switchboard for log data.
The core is written primarily in Ruby with performance-sensitive parts in C, giving a typical instance a footprint of roughly 30 to 40 megabytes of memory while processing around 13,000 events per second per core. Production deployments are packaged as fluent-package, the successor to the now end-of-life td-agent distribution.
Yes. Fluentd and its official plugins are fully open source under the Apache License 2.0, with no license fees or paid tiers controlled by the project itself.
Fluentd was created by Sadayuki Furuhashi, a co-founder of Treasure Data, with its first commit made on June 18, 2011, and the project open-sourced in October 2011.
Fluentd is written in Ruby and C with a large, community-driven plugin ecosystem best suited to complex routing and transformation, while Fluent Bit is written in C, embeds plugins into a single binary, and is optimized for high throughput in resource-constrained environments like containers, processing roughly 10 to 40 times more log volume on the same hardware.
Yes. Fluentd is a CNCF graduated project, a status it reached in April 2019, and continues to receive updates, with production deployments now recommended to use fluent-package, the successor to the retired td-agent distribution.
Fluentd itself does not sell a managed service or support contract. Paid support and training are available through the ecosystem, including a Linux Foundation training course and support from various logging and observability vendors.
Fluentd is governed as an independent, community-run CNCF project rather than owned by a single company. It originated at Treasure Data, which was later acquired by Arm in 2018 and spun out as an independent SoftBank-owned company in 2021, but that corporate history does not give any single company ownership of the Fluentd project.