Horizontal Pod Autoscaloing (HPA) governs the spinning and deletion of additional pods when the existing resources (CPU and Memory) of the microservice are exhausted. In Adeptia Connect, Autoscaling is by default disabled. You can enable HPA in Adeptia Connect by setting the required parameters in the global values.yaml file.
Follow the steps given below to configure HPA.
- Set the following parameters shown in the in the global values.yaml file.
Runtime HPA:
no of pf in queued state - shared/Autoscaling
Run time pod (RabbitMQ_Concurrency :10)
For shared queue - runtime
Autoscaliing: Threshold : 5
12 PF : 10 Running , 2 Queue
16PF : 10 running, 6: queue
For dedicated queue, we have option in UI.
Configuring HPA for runtime microservice before deployment (shared queue)
autoscaler>env:>min pod , and max pod, max queue is Threshold
Configuring HPA for runtime microservice after deployment (shared queue)
go to shared directory
Open AUTOSCALING file in the in the /shared directory in edit mode
2 = min pod
2 = max pod
5 = max queue or Threshold
performance = namespace
runtime = shared queue name
runtime = name of the deployment
Save the file.
Changes are reflected within 30 secs.
Configuring HPA for runtime microservice (dedicated queue) - done only after deployment
can be performed only by admin user.
Refere to creating a queue "Creating a queue - Adeptia Connect Help v4.0 - Adeptia Docs"