When creating a deployment you have several profiles to choose from, each with different scaling settings. This allows you to choose a theoretical maximum deployment cost.
The list below contains values only for web nodes that run PHP. Other resources, such as Redis or Varnish, are not included. Elastycloud will optimize and adapt these profiles over time to better suit the profile's purpose. Profiles may change to a small degree without prior notice. The requests and limits below are multiplied by the current number of required replicas, based on current load.
This is the profile with the lowest possible requirements for the store to be accessible. Only intended to be used when the store is not in use and for pre-launch testing where performance is not a requirement.
Auto-scaling is disabled for this profile. Resource starvation is a real possibility; please try switching to a different profile before contacting support.
Good for testing the platform, and limited load testing since it always has at least two replicas. However, because it has low pre-allocation it may be slow to get up to speed.
This is the lowest profile you should be running a real production store in. There is only one web node by default, but it is powerful and handles a fair amount of load with good response times.
Similar to small, but scales higher to allow for a larger concurrent load.
Scales even higher than the medium profile, allowing for a large amount of concurrent visitors.