The Research Computing team at The Alan Turing Institute is responsible for allocating, managing and supporting access to HPC and Cloud services for a large number of researchers and projects. We provide access to all kinds of compute resources, from targeted Tier 2 HPC systems through to fungible credits for Cloud services that can be used for any available resource.

Managing these resources efficiently can be challenging, especially when budget controls offered by Cloud service providers are limited with no way to prevent overspend by users. This presents serious challenges for adoption in research organisations like The Alan Turing Institute that need to centrally disseminate Cloud resources to researchers with independent budgets.

In this talk we will present RCTab (Research Computing Tables), our open-source budget and subscription control system for Azure which enables central management of Cloud resources while enforcing strict budget controls. We will discuss the background to the tool and demonstrate its functionality, including budget allocation, usage monitoring and shut-down of resources when the budget has been consumed.

We will also discuss our experiences developing the code in Python using the Azure REST API, including our experiences in converting the codebase of internally-used tools to a deployable, customisable and open source application with broader applicability through the use of the Pulumi Infrastructure as Code provisioning environment.

Links: 1. The RCTab documentation 2. RCTab source code main repo 3. RCTab CLI repo 4. RCTab API repo 5. RCTab Function app repo