This folder includes workloads used for testing performance models. We used a set of configurable benchmarks with different types of workload integrated in them. Thus, by modifying the requests made to each benchmark set, we can get a range of workloads that are a combination of different types of tasks common in serverless computing. Here is a list of functions used in this work:

• Bench1
• Autoscale-Go

NOTE: Please note that each of these functions include a combination of standard benchmarks used in previous studies. A combination of these benchmarks can represent different levels of stress on different resources of the virtual machines in the platform. You can find further details of these benchmarks in their dedicated pages.

Here is a list of workloads used in this study:

Experiment	Experimental Results	Model Prediction
ArrivalRate20

Experiment	Experimental Results	Model Prediction
ArrivalRate20

Experiment	Experimental Results	Model Prediction
ArrivalRate20

Experiment	Experimental Results	Model Prediction
ArrivalRate20

• “Shutdown Signals with Docker Entry-point Scripts” by Benjamin Cane
• Google CloudRun: Container Runtime Contract