NGF: document NginxProxy useClusterIP setting#2116
Conversation
|
can we target this to release branch for ngf-release-2.7 |
Document the new NginxProxy useClusterIP field, which routes upstream traffic to a Service's ClusterIP instead of individual Pod IPs to reduce NGINX reloads on backend Pod churn. Notes the fallback cases (headless, ExternalName, L4/stream) and the default value.
d65a783 to
1658297
Compare
|
Thanks @salonichf5! I've switched the base to |
|
I think we should mention the downside of using this is you lose out on NGINX's load balancing ability, and instead you end up just using k8s default Service load balancing. |
Enabling useClusterIP makes the upstream a single Service VIP, so NGINX load balancing across Pods no longer applies; the Kubernetes Service handles load balancing instead.
|
Thanks @sjberman! Added a note calling out the trade-off: enabling |
salonichf5
left a comment
There was a problem hiding this comment.
overall looks good, just a small comment
Co-authored-by: Saloni Choudhary <146118978+salonichf5@users.noreply.github.com>
Proposed changes
Problem: NGINX Gateway Fabric is adding a
useClusterIPfield to theNginxProxyresource (nginx/nginx-gateway-fabric#5557, resolving nginx/nginx-gateway-fabric#5554). The data plane configuration guide does not yet document it.Solution: Add a "Route upstream traffic to the Service ClusterIP" section to
content/ngf/how-to/data-plane-configuration.md. It explains what the setting does and why (fewer NGINX reloads on backend Pod churn, service-mesh compatibility), gives anNginxProxyYAML example, and documents the fallback cases (headless Services, ExternalName Services, L4/stream upstreams) and the default value.Testing: Rendered/reviewed the Markdown locally; no functional changes.
Related: nginx/nginx-gateway-fabric#5557, nginx/nginx-gateway-fabric#5554