Scale down is automatic. There is a scale down stabilization window of 300 seconds of waiting time before it goes from let's say 5 replicas to whatever lower number is needed, e.g. 3 replicas; so, when the condition are met for only 3 are replicas, it waits 300 seconds still holding the 5 replicas and then scales down to 3. Then later there is another 300 seconds of cool down period before it scales down to the minimum number of replicas (e.g. 0) when there is no need for more replicas than the minimum amount. BTW, the default scale up stabilization window is zero seconds, obviously :)
One of the best tutorials on Azure Container App! Thanks for these treasures.
Thanks bro your Tutorial is amazing
excellent, concise and clear video. Thanks
Awesome man .. this is great... Stuff... We are actually not aware of such amazing but simple videos... You are awesome
great man owsome job
good video-- very helpful. what load test tool did you use here??
One doubt I have is can we do custom api token validation in the container ingress? I mean custom API Authorization
once the replica is created, does it need to be manually removed if the request is getting low? thanks.
just from curiosity is there any example for scaling container app based on memory :)
You have to use keda metrics.
What about scale down? Can these server terminated if request goes down?
Scale down is automatic. There is a scale down stabilization window of 300 seconds of waiting time before it goes from let's say 5 replicas to whatever lower number is needed, e.g. 3 replicas; so, when the condition are met for only 3 are replicas, it waits 300 seconds still holding the 5 replicas and then scales down to 3. Then later there is another 300 seconds of cool down period before it scales down to the minimum number of replicas (e.g. 0) when there is no need for more replicas than the minimum amount. BTW, the default scale up stabilization window is zero seconds, obviously :)
Normally there should be 20 replicas after running test with 400 threads and 20 concurrent requests for the scaling rule?
3:19 He set the maximum number of replicas to 5.