This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(118.4, 87%, 21%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EM%3C/text%3E%3C/svg%3E)
I have a function app that's successfully running fine. However I am trying to set rules to create a new instance when CPU usage is above 90%. In such case, how does load balancing works between two instances?
Is there any inbuilt options for this or should I use any other application to handle load balancing effectively?
Hi Man-2022
Have you considered using a probe for this?
"HTTP / HTTPS probes can be useful to implement your own logic to remove instances from load balancer if the probe port is also the listener for the service. For example, you might decide to remove an instance if it's above 90% CPU and return a non-200 HTTP status."
Thanks @David Broggy for your inputs. I am new to Azure so trying to first clarify few terms.
I don't see any application called load balancing inside the function app. So when you say "Load Balancer" do you mean adding Azure Load Balancing to my function app?
or even something like Azure front door?
I read about probes in short, but I don't want to get non-200 status code because of the load on the active instance of function app.
Since HTTP trigger can create a new instance very quick, I only want to create a new instance when CPU goes beyond certain limit say above 85-90% or so.
In such case does Azure function has an internal balancer within them to divert the triggers to the newly assigned instance ?
or
Does the new triggers are automatically routed to new instances until the CPU in the old instance come down below the defined limits?
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(275.2, 33%, 36%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EMM%3C/text%3E%3C/svg%3E)
Hi @Man ,
Thanks for reaching out to Q&A forum.
It appears that you are trying to set up autoscale rules based on CPU utilization metrics. Usually the requests are routed to instances based on round robin method. In your case, when the Function app is running on one instance and CPU utilization is more than 90%, then a new instance gets added. Initially, the requests will be routed to the new instance as the previous instance is exhausted. Basically, the load balancing is handled by the Function app infrastructure and there is no need to configure load balancing exclusively.
Hope this helps! Feel free to reach out to me if you have any questions or concerns.
Please 'Accept as answer' and ‘Upvote’ if it helped so that it can help others in the community looking for help on similar topics.
Hello @MughundhanRaveendran-MSFT ,
Thanks for your comments.
If I understand your statement then all existing requests will be handled within the 1st instances which is already running and then the instance will be release.
Meanwhile all the new requests received will be routed to the 2nd instance (new instance created) once its available.
If the above said statement is correct then if I schedule to create multiple new instance based on any condition will the requests be handled between each instance in some rational order?
@Man , Yes. the App service Front end will route the requests in a rational manner