Hi
I'm possibly doing something wrong, so I'm open to suggestions here.
I have a pool of application servers ( linux virtual machines), an Azure Mysql, and an Azure Redis.
We have been having occasional network issues between the machines in the pool, between the machines and Azure Mysql, and between the machines and Redis.
I'd like to set up Connection Monitor to track / monitor this, but I would like to do it via Infrastructure as Code (ARM )
I've managed to create an arm template that makes a Connection monitor ( not the simplest job, the example given relies on you figuring out you need to create it as a child of the NetworkWatcherRG and you need to create it under the NetworkWatcher called Networkwatcher_<location> ) but had to hard code some things, and am having trouble figuring out how I would do this using a copyIndex to actually set up Connection Monitors using all of our application servers.
So:
- is this even theoretically possible ? What I was initially thinking was I would create the ConnectionMonitor with only the endpoints for the mysql and redis, and then create a second subresource for endpoints for the application servers which used the copyIndex() and another subresource for the tests again using a copyindex. but when I tried deploying the ConnectionMonitor without any tests it complained. what I'm currently thinking I'm going to have to do is is put app-0 in the initial ConnectionMonitor, and then do the sub resource thing with the rest of the copyindex, but wondering if that's even going to work.
- we're in the middle of moving our app servers to scalesets. do i have to put something in the machine image that adds itself to the connectionMonitor? is there a way to do this automatically?