Fluentd

Fluentd is the CNCF-standard log collector that became infrastructure's duct tape. With 1,000+ plugins, it connects virtually any log source to any destination. It's the "universal logging layer" that sits between your apps and whatever observability backend you picked this year.

If you're in a Kubernetes shop that values CNCF-standard tooling and vendor neutrality, Fluentd is the safe choice. Vector (Rust, Datadog-backed) is 5-10x faster with lower memory and a better configuration language — it's what you'd pick if starting fresh. Fluent Bit is Fluentd's lighter sibling, purpose-built for edge and container sidecar deployments. Logstash is the Elastic-coupled option.

The catch: Fluentd is showing its age. The Ruby core means performance lags behind Rust-based alternatives, and recent benchmarks show it dropping logs before hitting 10K logs per second under pressure. The plugin ecosystem is vast but uneven — quality varies wildly between maintained and abandoned plugins. For new deployments, Vector or Fluent Bit are objectively better choices.