Spike Arrest policy with Apigee edge microgateway on API Proxy level

[pallavjoshi]

Spike arrest policy for apigee edge microgateway works at microgateway(version 2.5.x) level. We would like to implement spike arrest per API Proxy, this is something which look possible with 2.6.1-alpha version c.f https://community.apigee.com/articles/46408/microgateway-plugins-per-proxy.html

Within current setup version 2.5, we may choose to have separate Microgateway cluster(loadbalancer with failover setup) for each API Proxy as workaround. How we can manage these multiple cluster within same organization.

[Former Community Member]

Currently, there is no way to implement spikearrest on each proxy.

[pallavjoshi]

Hello @srinandans thanks for your quick response. So as i mentioned workaround, having multiple cluster of microgateway within same organization is not possible? Also we will like to know when 2.6 is planned to be released.

[Former Community Member]

You can definitely have multiple MG to the same org/env. They can even run on the same VM.

There are no plans to release MG 2.6.1 at this time.

[pallavjoshi]

Okay, Please note here we are discussing possibility of running different micogateway cluster. So as a workaround for spike arrest limitation if we try to setup multiple microgateway clusters (one for each API Proxy) all having unique DNS, how we can manage it within organization. Can we use multiple virtual_host in organization for this purpose or what would be best recommended way. @srinandans ( tag added ;))

[Former Community Member]

hmm. This requires a meeting and I'd rather not speculate on design here on community.

Running a MG per API (in what we call a sidecar pattern) is quite common when using Kubernetes or Cloud Foundry. K8s and CF already provide a unique hostname for every API. I'm assuming you're not using either. I suppose something similar may be done on VMs too.