k8sgpt/pkg/server

serve

The serve commands allow you to run k8sgpt in a grpc server mode. This would be enabled typically through k8sgpt serve and is how the in-cluster k8sgpt deployment functions when managed by the k8sgpt-operator

The grpc interface that is served is hosted on buf and the repository for this is here

grpcurl

A fantastic tool for local debugging and development is grpcurl It allows you to form curl like requests that are http2 e.g.

grpcurl -plaintext -d '{"namespace": "k8sgpt", "explain" : "true"}' localhost:8080 schema.v1.ServiceAnalyzeService/Analyze

grpcurl -plaintext  localhost:8080 schema.v1.ServiceConfigService/ListIntegrations
{
  "integrations": [
    "prometheus"
  ]
}

grpcurl -plaintext -d '{"integrations":{"prometheus":{"enabled":"true","namespace":"default","skipInstall":"false"}}}' localhost:8080 schema.v1.ServiceConfigService/AddConfig