Cloud foundry scale

In the elastic runtime, the default settings for a container are set. These include the memory, ephemeral disk etc. As part of the cf push (either thru manifest file, or at command line) you can override those parameters. With push of an app, the LRP settings and the archives are uploaded to the cloud foundry blob store.

Scaling can be vertical (increase memory, and/or disk) or horizontal (additional instances). Also scaling can be manual or auto-scaling (more on it later).

My understanding is that, in both cases of manual scaling, cloud foundry will create a new container instance with the new values (mem, disk). CF will pull archives for the LRP from the blob store, and push it to the container and invoke the LRP. Once it up, the old instance is shutdown and reclaimed.

Autoscaling is for horizontal scaling only. First you need to create a service instance of the auto-scaler. You bind the auto-scaler with your LRP and provide details on how the auto-scaler should work - cpu %, min, max, number of instances etc. Once that has been set, CF will auto scale the app/LRPs based on the provided details.

I am no expert in cloud foundry. So, I request the experts to please correct me. It is always good to learn.

Hope this helps!