sticky-session | Sticky session balancer based on a cluster module

I am using two ec2 instances on a classic load balancer. My users were having 502 errors in the middle of their analysis, especially while having ajax calls. In other words, if a user is connected to a server instance, an ajax call in the middle of using the program is directed to the other server instance. And that causes that error. I need the user targeted to the same server instance during the his/her session.

After googling the issue mentioned above, I found this:

Load Balancing, AJAX, and You

https://docs.aws.amazon.com/elasticloadbalancing/latest/application/sticky-sessions.html

According to these documentations, I should apply sticky sessions to my load balancer. However, it requires clients to have their cookies enabled on their browsers. I really don't want that.

Is there any other solution for this issue? I really would appreciate any suggestion.

I have the following configuration for request-based sticky session using Spring Cloud LoadBalancer

We are have enabled horizontal pod autoscaling in GKE, our pods are sitting behind a clusterIP type service and we are routing public traffic to that Service using NGINX Ingress controller. When monitoring the usages we have noticed that traffic is not equally distributed between pods. it's routing traffic to one single pod. but whenever we manually deleted that particular pod it will route traffic to another available pod.

Is there any way we can enable ingress rules to distribute traffic equally

Ingress

Using spring-boot 2.17 with hazelcast 3.11 and tomcat 9.0.39.

I have set a WebFilter bean:

I am trying to implement sticky sessions. I have found out that I can add a prefix to the exiting cookies servers to identify them. See here for more

I have a deployment which comprises 2 pods with a webapp in them. Nginx ingress controller is installed to expose those 2 pods using tls-passthrough. I have tried to use annotations to setup sticky-sessions but to no avail. No cookie is present in the headers hence no stickyness. What could I do to make it work ? Many thanks for your wise answers.

The Ingress I have created :

I have created a Jmeter script to check the performance of a site. The website is hosted in AWS with elastic scaling and with sticky sessions. Basically the AWS load balancer will assign a session cookie to each user so the load balancer can direct the user to the correct instance.

source

My problem is, as I'm using a cookie manager and clearing all the cookies with each iteration. Does it clear these assigned cookies too? I suspect this because the script error rate is lower when we execute the script on a single AWS instance than in auto-scaling ( multiple instances )

Any idea ?

I have a Flask app running in AWS using Flask-JWT-Extended. It is serving REST API calls to a web app.

As I understand from the documentation, the tokens are generated and stored in memory. I am considering storing them external to the Flask app in either a database or Redis. The reason for this is to support load balancing:

• I presume that sticky-sessions would be required to make sure that the client's token can be properly decoded and analyzed for validity
• I am considering putting the app in AWS Lambda, which would probably wipe out the JWT list once the request was served.

My questions are:

• Is there any reason this scheme would not generally work?
• If the tokens are stored outside the Flask app, it is not clear how to override the local token storage and access an external storage medium. Can this be done?