I deleted accidentally the ApiGee instance

[jebadia]

Hello.

I am new at ApiGee. And I was deleting the ENV to minimize the cost for my company. But doing so, I accidentally deleted de ApiGee Instance. 

After that I hurried and created a new one. But now, every time I do an Api Call, Postman will return:

I went to the load balancer to see if there is a problem and I found this:

When I click in that link, the Compute Engine page is shown to me with this info:

Is like everything if fine but at the same time is not.

Please can anyone help me with this. If there isn't a straight solution, please could anyone tell me how restart APIGee to default or delete everything and start from scratch.

[hartmann]

Hello, 

If I were to guess, your Apigee instance IP changed, meaning you would need to reflect this change within your Managed Instance Group/Instance template. To do so is as follows:

1. Copy the previously used instance template (under virtual machines) via 'create similar' under the instance template blade to essentially copy/paste the current template. From there, you can navigate to Custom Metadata -> ENDPOINT and modify the endpoint to point to your new instance (which can be found under Apigee -> Management -> Instances -> IP address)
2. Now that the new instance template is created, you can add either edit the instance group to point to the new instance template (pointing to the new IP address), or create a net new instance group to point your LB to/through
3. Depending on the above, you would then need to (potentially) modify your load balancer to reflect these changes as needed (ie: create a new backend service to point to the updated instance group set)

The following should be helpful in this situation as well: https://www.googlecloudcommunity.com/gc/Cloud-Product-Articles/How-to-upgrade-Apigee-MIGs-without-di...

Let me know if you have any additional questions and/or concerns - thanks!

[jebadia]

Hi, I did it as you say, but the Load Balancer health check keeps failing:

This is the default configuration for health check is:

[hartmann]

Hello,

Are the virtual machines backing the underlying Managed Instance Group is online? Additionally is there any reason why you are health checking over port 80 - most customers I have seen typically use port 443/HTTPS (which should work to expectation)

Thanks!

[jebadia]

Yes, the VM is up and running:

For the health check, I have changed between those ports, but in both, the error is still present.

[hartmann · Accepted solution]

Hello,

Thinking this through it might be easier if we ported your workflow from MIGs to PSC-driven (given the ease of use/simpler access pattern to the Apigee X runtime). I created the following video (noted here: https://www.youtube.com/watch?v=gNqTDcvcEL4) which covers the reasoning behind the change as well as a step by step as per how to migrate inline

Let me know your thoughts and if this resolves the issue at hand (which it should assuming you have the proper Apigee instance IP of the new runtime/the instance is attached to the new environment) - thanks!

[jebadia]

Ok, thanks I will comment you.

[jebadia]

Thanks you so much. This video save my life.