Fluent Bit is the tiny, fast log shipper you will find running on almost every Kubernetes cluster.
Fluent Bit is a lightweight, high-performance open-source (CNCF) log and metrics processor and forwarder. It collects logs from files, containers and other sources, parses and filters them, and ships them to a central destination, with a very small memory and CPU footprint.
Fluent Bit sits close to where logs are produced and does three things fast: collect (tail log files, read container output, receive syslog), process (parse, filter, redact and enrich), and forward (ship to a backend or another collector). Written in C, it is built for a tiny footprint and high throughput, which is why it is popular as a per-node agent where resource use matters. It is the successor-in-spirit to Fluentd, trading some flexibility for much lower overhead.
Fluent Bit's most common role is as the log-shipping DaemonSet on Kubernetes clusters: one lightweight instance per node, tailing container logs and forwarding them onward. Its small footprint makes it ideal for this at scale. In modern setups, Fluent Bit increasingly ships logs into an OpenTelemetry Collector or directly over OTLP, so it acts as the fast edge collector feeding a unified telemetry pipeline.
A Fluent Bit configuration is built from a small set of stages that mirror what it does. Inputs collect data, tailing container log files, reading systemd or syslog. Parsers turn raw lines into structured records. Filters modify records in flight, adding Kubernetes metadata, dropping noise, redacting sensitive fields. And outputs forward the result to one or more destinations. This input, parse, filter, output flow is what makes Fluent Bit both fast and flexible: you shape and reduce logs at the edge, close to where they are produced, before paying to ship and store them.
As OpenTelemetry has become the standard, Fluent Bit's role has evolved rather than disappeared. It remains the efficient edge collector for logs, but increasingly it forwards those logs over OTLP into an OpenTelemetry Collector that unifies them with traces and metrics. This combination plays to each tool's strength: Fluent Bit's tiny footprint and speed at the node, and the Collector's vendor-neutral, all-signal pipeline downstream. Understanding Fluent Bit as the fast front end to an OpenTelemetry pipeline, not a competitor to it, is the modern way to place it.
Fluent Bit-collected logs flow naturally into Ops Singularity over OpenTelemetry, so teams keep their fast edge log shipper and add TelemetryOps and Sentinel AI to correlate and resolve on top of what it forwards.
Collecting, parsing and forwarding logs (and some metrics) with a very small footprint, most commonly as the per-node log shipper on Kubernetes.
Both are CNCF log processors; Fluent Bit is lighter and faster with a smaller footprint, while Fluentd is more feature-rich and plugin-heavy. Fluent Bit is preferred as a lightweight edge agent.
Ops Singularity turns open telemetry into autonomous, governed resolution. See it on your own stack.