# Does port 8000 run on head only or both workers and head

**URL:** <https://discuss.ray.io/t/does-port-8000-run-on-head-only-or-both-workers-and-head/22729>\
**Category:** Ray Serve\
**Created:** [June 24, 2025, 3:40pm UTC](https://discuss.ray.io/t/does-port-8000-run-on-head-only-or-both-workers-and-head/22729 "2025-06-24T15:40:06Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![mirage](https://avatars.discourse-cdn.com/v4/letter/m/eb9ed0/32.png) [@mirage](https://discuss.ray.io/u/mirage)\
**Post date:** [June 24, 2025, 3:40pm UTC](https://discuss.ray.io/t/does-port-8000-run-on-head-only-or-both-workers-and-head/22729/1 "2025-06-24T15:40:06Z")

</div>

Using Ray 2.46 in EKS  
I got 2 pods Head and Worker.  
If i exec into pod and check listening ports then in one example i see port 8000 in both Head and worker.

and in other case port 8000 was not on worker.

So health check failed as nothing is listening on port 8000.

I have not configured readiness probes in worker but they still get inserted into by by opertaor i think

---

<div class="post-metadata">

**Author:** ![christina](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/christina/32/7542_2.png) [@christina](https://discuss.ray.io/u/christina)\
**Post date:** [June 25, 2025, 10:53pm UTC](https://discuss.ray.io/t/does-port-8000-run-on-head-only-or-both-workers-and-head/22729/2 "2025-06-25T22:53:35Z")

</div>

Hello, they mention this a bit in the docs here: [RayService worker Pods aren’t ready — Ray 2.47.1](https://docs.ray.io/en/latest/cluster/kubernetes/user-guides/rayservice-no-ray-serve-replica.html#codecell4) and here: [RayService worker Pods aren’t ready — Ray 2.47.1](https://docs.ray.io/en/latest/cluster/kubernetes/user-guides/rayservice-no-ray-serve-replica.html#toctree-checkbox-70)

But essentially:

> Look at the output of Step 4.2. One worker Pod is running and ready, while the other is running but not ready.  
> Starting from Ray 2.8, a Ray worker Pod that doesn’t have any Ray Serve replica won’t have a Proxy actor.  
> Starting from KubeRay v1.1.0, KubeRay adds a readiness probe to every worker Pod’s Ray container to check if the worker Pod has a Proxy actor or not.  
> If the worker Pod lacks a Proxy actor, the readiness probe fails, rendering the worker Pod unready, and thus, it doesn’t receive any traffic.

Essentially, only worker pods running a Ray Serve replica will have port 8000 open and be marked as ready, do you have Ray Serve running on the other worker confirmed?