Cold Start Test

The project consists in the implementation of a system able to monitor the performance of an external API. As the API is implemented using serverless, the system tests the performance in two main scenarios:

1. Cold Start: rare/infrequent calls are made to the API, thus awakening the sleeping serverless micro-services;
2. Warm Start: frequent calls are made to the API, thus testing the “normal” performance.

The goals of the project are multiple:

1. Test the external system;
2. Practice how to implement a NoSQL database using Azure Cosmos DB;
3. Practice how to implement a BI dashboard using Power BI.

The overall architecture is composed by three main blocks:

1. Main platform implemented on Microsoft Azure;
2. REST API backend;
3. Client/Laptop for serving data visualisation dashboard.

The choices for reaching the design are:

Here below can be seen a detailed view of the Microsoft Azure implementation. The architectural flow is not displayed as it is quite straightforward:

1. Azure Function wakes up periodically (with changing period based on cold/warm mode).
2. Azure Function invokes the REST API backend, measuring outcome and performance.
3. Azure Function stores the data in a SQL-API Cosmos DB database.
4. Power BI retrieves the data from Cosmos DB and displays them in a dashboard.

Here below there are the screenshots related to the Power BI dashboard.