Scale your Cluster

Scaling options

Hydrolix provides flexible scaling to accommodate your changing infrastructure and usage needs.

Scale by component: Scale a component's resources or replica count in the scale section of your hydrolixcluster.yaml file.

Stateful vs. Stateless scaling: Hydrolix supports both stateful and stateless components, and scaling these requires different considerations.

Scale profiles: Scale all components in a cluster using the pre-made Hydrolix scale profiles that are designed for different infrastructure and throughput levels. For example, scale_profile: prod provides components scaled for a typical 1-4 TB/daily ingest workload.

Scale profile overrides: Override a component’s scale profile by setting the scale field in hydrolixcluster.yaml.

Scheduled overrides: Horizontally or vertically scale components based on a weekly schedule, a time window, or cron schedule.

Scale to zero: Scale all replicas in the cluster to zero except for the operator pod.

Scale to minimal: Scale most components to zero while leaving the cluster available for API calls and the UI.

Custom autoscaling with Prometheus metrics: Autoscale components in a cluster based on Prometheus metric values.

Overcommit: Ignore Kubernetes CPU and memory requests and limits for Kubernetes containers. Useful when running in a resource-constrained environment.