I'd really have to sit down and think about it, but I think it'll be difficult. Your agents are hard-coded to authenticate themselves to a specific Access Key and Controller Name. Dynamically creating new Controllers would require you to also dynamically create new controller-info.xml files. Also, license key management might be an issue because the license rules are hard-coded into the key. That's not to say it's impossible, but you'd have to overcome those, perhaps a shared .yml like they do with the cluster agents? Not sure.
You can probably definitely do it for the Events Services and maybe even the EUM server since those can scale pretty easily. But you'd have to take in some extra considerations on the Console for the Events Server and maybe a reverse proxy to handle a dynamics EUM servername.
Experts,
Any thoughts on this?
Hello experts,
Just out of curiousity and with intermediate knowledge on kubernetes, I was just wondering if anyone tried installing Appdynamics controllers on Kubernetes or some other docker based platforms? And if yes, then what are his observations in terms of scalability and performance.
Reason am interested in such a kind is - we all know there is a constraint within AppD wrt horizontal scaling of controllers (we cannot have clustered controllers within active leg of architecture), I was thinking to create a base image for AppD controllers(standalone) and as the demand increases, you spinup another controller with no delay.
I understand that this will increase complexity of too many controller info to manage, but atleast time consumed in building of new clustered AppD env will be reduced (and cheaper than having full physical/virtual box provisioned)
I might sound weird but thought to highlight it over here for comments and suggestion.
(Like AppD on aws is awesome concept but due to some restrictions am not able to proceed on that thread)
Appreciate your help.
Regards,
VVS
