Kind Clusters¶
Atlas keeps Kind cluster definitions under ops/stack/kind/ so local and CI
cluster shape stays declared and reviewable.
flowchart TD
Kind[Kind cluster variants] --> Small[small]
Kind --> Normal[normal]
Kind --> Dev[dev]
Kind --> Perf[perf]
Small --> Quick[Quick validation]
Normal --> Baseline[Baseline local simulation]
Dev --> Iteration[Developer-focused iteration]
Perf --> Load[Heavier validation and load]
These variants exist because Atlas does not pretend one cluster shape can serve every validation goal honestly. The Kind definitions let operators choose a cluster capacity and port shape that matches the profile and the suites they intend to run.
Cluster Variants¶
cluster.yamlcluster-dev.yamlcluster-small.yamlcluster-perf.yaml
How They Differ¶
cluster-small.yamlreducesmax-podsto60and is the right fit for smaller local or CI profilescluster.yamlandcluster-dev.yamlusemax-pods: 110for baseline and developer-oriented cluster simulationcluster-perf.yamlraisesmax-podsto220for heavier profile pressureops/k8s/kind/cluster.yamlis the simulation-focused Kubernetes cluster definition and should be read alongside, not instead of, the stack variants