// the take
Loggie is a Go-based log shipper and aggregator built around Kubernetes CRDs (LogConfig, ClusterLogConfig, Interceptor, Sink) instead of static config files. It's aimed at platform teams running Kubernetes who want a Filebeat/Fluent Bit replacement with declarative, GitOps-friendly pipeline management and built-in alerting.
Multiple independent pipelines with multiple outputs per pipeline, unlike Filebeat's single-queue model — useful if you're fanning logs out to several sinks with different processing. CRD-driven config delivery means pipeline changes propagate through kubectl/GitOps instead of SSH-ing config files onto nodes. It folds log-based alerting and metric aggregation (count/sum/avg by groupBy over a log stream) directly into the interceptor chain, which normally requires bolting on a separate exporter or stream processor. Adaptive sink concurrency that backs off based on downstream response is a real feature, not just marketing — most shippers only do static batch/worker tuning.
No commits since mid-2024 per the push date, which is a real risk for something you'd run as a DaemonSet in production — check the issue tracker for signs of abandonment before adopting. The performance numbers in the README are a single self-run benchmark (one file size, one Kafka setup, no percentile/latency breakdown, no comparison against Fluent Bit or Vector, which are the actual competitors here, not Filebeat/Logstash). Primary docs and community are Chinese-first (README_cn.md, NetEase origins show through) — expect rougher English docs and a much smaller pool of English-language troubleshooting threads. At 1.3k stars it's a tiny fraction of Fluent Bit or Vector's adoption, so you're trading ecosystem maturity and integration breadth for the CRD/alerting features, and you'll be debugging edge cases mostly on your own.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →