Certain improvements that I feel should be made would start with the UI perspective. The UI is not user-friendly, and newcomers or new users might get confused at times about how to proceed and what to do. This is an area AutoRABIT Automated Release Management (ARM) should work on. Additionally, the build and deployment job configuration is very powerful but complex. There are many settings, and the consequences of misconfiguration are not always clearly communicated until a deployment fails. We have also experienced occasional performance issues with the web application during periods of high concurrent activities, particularly when multiple teams are running deployments simultaneously. During those times, we feel lag. Regarding needed improvements, deployment progress visibility when we kick off a large metadata deployment involving hundreds of components is quite limited. There are times when it is genuinely difficult to tell whether the deployment is progressing normally or has silently stalled somewhere in the middle. A more granular real-time progress view would reduce a lot of anxiety during critical deployment windows. I would also mention the observability integration. We heavily use DataDog for monitoring our systems, and there is no clean native integration between AutoRABIT Automated Release Management (ARM) and DataDog. This means deployment events are completely disconnected from our application performance metrics, so correlating a deployment with a spike in errors requires manual cross-referencing between two separate tools. A proper webhook or native integration would be a significant improvement.
