Logging Operator is a Kubernetes operator from Kube Logging (formerly Banzai Cloud) that manages Fluent Bit and Fluentd deployments declaratively. Instead of hand-writing Fluent Bit input plugins and Fluentd parser blocks, users describe logging pipelines as Kubernetes custom resources and the operator renders the configuration.
Four CRDs do the work. Logging defines a tenant and points at a Fluent Bit DaemonSet and a Fluentd StatefulSet. Flow and ClusterFlow select which pods’ logs to route via label selectors and declare filter and parser chains. Output and ClusterOutput describe destinations — Loki, Elasticsearch, Kafka, S3, Splunk, Datadog, and more — complete with authentication from Kubernetes secrets. The operator generates the matching Fluent Bit and Fluentd configs, reloads the pods, and handles rolling upgrades.
Logging Operator is a CNCF sandbox project and the log-collection layer behind Rancher’s fleet logging, Kubesphere, and several managed Kubernetes platforms. It competes with the OpenTelemetry Collector operator and with hand-rolled Helm charts for Fluent Bit, and is the right pick when teams want multi-tenant log routing expressed as Kubernetes objects rather than ConfigMaps.