NVIDIA / NVIDIA/nvcf

Standardize health endpoint path and response across services

Open
#334 0 comments 0 reactions 1 assignee View on GitHub

@priyaselvaganesan is already working on this.

Since Aug 31, 2026.

Dominant language
Go
Stars
218
Forks
72
Avg merge
1d 12h
Merged PRs (30d)
427

Description

Description

Services in this repository use inconsistent paths and response formats for their health endpoints.

Service Plane Path Response
grpc-proxy Invocation /health JSON with dependency status (NVCF API gRPC, NATS) via hellofresh/health-go
vanity-gateway Invocation /health JSON with dependency status (NVCF API HTTP) via hellofresh/health-go
ratelimiter Invocation /health 200 or 500 based on Olric cluster health
http-invocation Invocation /health unknown
llm-api-gateway Invocation /healthz 200 empty body (unconditional)
helm-reval Control /healthz 200 OK body "OK" (plain text, unconditional)
nats-auth-callout Control /healthz + /health JSON {"status":"ok"} (unconditional)
function-autoscaler Control /health, /admin/health/liveness, /admin/health/readiness JSON with component health status
nvcf-unbound Compute /health 200 OK body "ok" (plain text, unconditional)
nvca Compute /healthz 200 via go-lib core framework
nvsnap Compute /health (main + agent), /healthz (webhook), /v1/healthz (blobstore) varies across three internal servers
worker-task Compute /v1/health/live + /v1/health/ready via go-lib nvkit
worker-utils Compute /v1/health/live + /v1/health/ready JSON via hellofresh/health-go

Note: byoo-otel-collector, ess-agent, image-credential-helper, worker-init, and worker-llm-credentials have no self-health HTTP endpoint found in this repo.

Definition of Done

  • All services use a single consistent health endpoint path
  • The path is registered as the standard in go-lib
  • Helm chart liveness and readiness probe configurations reflect the chosen path
  • Response format consistency across services needs further triaging — some services perform real dependency checks while others return unconditional 200 OK, which may be intentional or may represent gaps in health signal coverage

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.