Adding a new backend to load balance, stuck at "Fetching health status"

[Jason Lee]

Hi, I'm in desperate need of help.

I'm trying to add a new instance group to the load balancer where it already has one instance group.

But this is a new instance group in a different zone so it requires a *new* backend.

Everything seems ok, all up to the point where I finalized and save -- at which point the loading wheel spins forever and says "fetching health status". This leads me to believe the health checks are not passing and thus new backend cannot be created.

Can someone please help me why and how to solve this?

[Nur]

Could you provide more details on your setup? Are you using HTTP LB with NEG (Network EndPoint Groups) backend? if so, we already have reported issue where the backend constantly showing "fetching health status". The product team already working on this issue and no ETA is available at this moment.

[Jason Lee]

I'm using a HTTP(S) LB. Not sure if I'm using NEG. If you let me know how I can check if I'm using NEG, I can let you know.

[Nur]

A Network Endpoint Group (NEG) for Load Balancing is a collection of Network Endpoints (An endpoint device is an Internet-capable computer hardware device on a TCP/IP network) where each network endpoint is an IP:port pair. They can now be used as a target for load balancing, allowing you to accurately load balance your service traffic across applications or containers hosted inside your VMs. 

[1] gcloud beta compute network-endpoint-groups: https://cloud.google.com/sdk/gcloud/reference/beta/compute/network-endpoint-groups/

[Mohammad Nurul Islam]

Following up on my previous message,  to check if you are using NEG or not you can use this command 

$ gcloud beta compute network-endpoint-groups list

For more details on this command and NEG review this GCP documentation [1].

[1] gcloud beta compute network-endpoint-groups: https://cloud.google.com/sdk/gcloud/reference/beta/compute/network-endpoint-groups/

[Michael Gardner]

We're getting this issue too, but we're not using NEG. We've got an HTTP(S) load balancer with one backend service, backed by a single GCE instance group. Trying to add a new backend service using the web console results in a "Fetching health status" message on the finalize step that never completes.

[Steven De Coeyer]

Having the same issue. Anyone have a clue?

[Mahmoud Sharif (Google Cloud Platform)]

This issue is under investigation by the product team. This is related to  to the fact that HTTP Load Balancer's NEG backend is spinning with a "Fetching Health Status" message. If you are not using a NEG, I invite to open  a new Issue in the Public Issue tracker [1][2].

  

 [1] https://developers.google.com/issue-tracker/

 [2] https://developers.google.com/issue-tracker/guides/access-ui