0

In my current application server which currently runs not in a Kubernetes environment, we use IPS in order to communicate with a specific instance of the same application agent.

the application is stateful and therefore we have to communicate each time with the same instance, with IP as an identifier.

Now we want to deploy this agent application to Kubernetes, however, because it's a stateful application I am a bit confused. How can I communicate each time I start a session with the same instance application? Can I have an IP or some another identifier of the pod? how will it work with the Ingress resource?

I looked at the Statefulset object but also did not understand how can it help me. I also looked at "sticky session", but I don't want to define an amount of time in order to define one session. I want to be able to access the same pod, and if the pod dies I want to get an exception.

How can I do it in Kubernetes? is it possible?

Ohad
  • 1,563
  • 2
  • 20
  • 44
  • To understand your requirement further : Have you deployed your application in 2 instances and established a communication between the applications without Kubernetes? Now Do you want to deploy the same application in 2 PODs and want to establish communication? Do you have any plans to scale your application? If yes, how do you want to communicate in such a case? – Hemanth Kumar Jan 12 '23 at 15:21
  • Hemanth Kumar - 1. yes today my application run outside Kubernetes. its one main server that communicate with 100 or more agent applications using IP's as identifier. 2. my initial plan is to keep the server outside of the cluster and just the agents applications keep in the clusters – Ohad Jan 15 '23 at 08:25
  • Can i know which cloud environment you are using? This will vary from one cloud environment to other. – Hemanth Kumar Jan 23 '23 at 12:28
  • on-premise Hemanth Kumar – Ohad Jan 23 '23 at 14:21
  • It is not possible to directly receive the alert when POD dies in accordance with your request. Because you are using on-premises, you will need to use a monitoring tool like Promotheus. For more information, see this [SO](https://stackoverflow.com/questions/69985164/alerts-in-k8s-for-pod-failing) – Hemanth Kumar Jan 24 '23 at 07:11

0 Answers0