I developed an application, that makes some calculations, and those calculations consumes memory, the user need to use same machine recursivelly then after 15 minutes without use the memory is cleared, but if I got many users acessing the instance I got memory out problems, its normal, because the algorithm was created to consume memory, the point is I can tell the autoscaling to scale if processing goes above 80%, but I can’t scale if memory goes above 80%, anyone know or have an idea about how to scale GAE Standard based on memory consumption?
TIA
Advertisement
Answer
The only possible automatic scaling settings that can be configured for the App Engine standard environment are:
- target_cpu_utilization
- target_throughput_utilization
- max_concurrent_requests
Find the documentation here and these are the corresponding parameters that you’ll need to configure on your app.yaml file.
Notice that if you could somehow correlate the RAM used for your specific application with the concurrent requests, throughput utilization or the CPU utilization you could change the respective parameter(s) on your app.yaml file in order for your application to scale accordingly, but there is currently not a way to be able to enable automatic scaling on App Engine applications in the Standard environment based precisely on memory usage.