SignifAI's collection philosophy is API-driven. We believe it is much easier and more standard to integrate with systems over remote APIs. However, sometimes that is not doable, and we still want to be able to support different systems and applications. Because of that, we developed a set of collection and processing capabilities using an agent.
We spent some time in evaluating different agents (mostly open sourced) as well as developing our own agent. During the evaluation we have asked ourselves how can we achieve a single extensible flow that starts with data collection at all layers, then transform the data with basic filtering or adding information for context, and finishing with publishing the processed data to the SignifAI platform.
Eventually, we came to the conclusion that it makes more sense to collaborate with a solid agent framework which we believe is the closest to what we had built ourselves.
That is exactly the reason why we picked the Snap Telemetry Framework and extended it.