Skip to content

Scale Profiles

Overview⚓︎

Scale profiles provide predefined resource and replica settings for all cluster components. They give you a consistent baseline without tuning each service individually.

When to use a scale profile⚓︎

  • Create a new cluster. Start with a profile that matches your expected ingest and query load.
  • Standardize resources. Apply the same profile across environments for consistency.
  • Simplify scaling. Use a single profile instead of setting replicas for every service.

How scale profiles work⚓︎

Each profile defines default CPU, memory, and replica values for services in a cluster. You apply the profile once, and the operator propagates those settings across the cluster.

Scale profile changes apply cluster-wide.

Profiles are named for typical use cases. For example:

  • prod: tuned for steady ingest of 1–4 TB per day with balanced query load.
  • dev: lighter settings for development or test environments.

Set a profile⚓︎

Add the profile to your hydrolixcluster.yaml file:

1
2
spec:
  scale_profile: prod

The operator applies the prod settings to all cluster services.

Default profile⚓︎

If no scale profile is set, the cluster defaults to eval.

Available profiles⚓︎

Hydrolix includes predefined profiles for common use cases:

  • dev: light settings for small clusters and testing.
  • eval: evaluation settings for trying out features.
  • prod: production-ready profile for 1–4 TB/day ingest with balanced query load.
  • mega: large-scale production profile for 10–50 TB/day ingest.

Profiles provide a baseline and can be customized or overridden.

Node size recommendations⚓︎

16 vCPU nodes

For eval and prod deployments, use 16 vCPU nodes:

  • EKS: c5n.4xlarge
  • GKE: n2-standard-16
  • LKE: Dedicated 32 GB

32 vCPU nodes

For mega deployments, use 32 vCPU nodes:

  • EKS: c5n.9xlarge
  • GKE: n2-standard-32
  • LKE: Dedicated 64 GB