Alright, I've been hammering on Granola for a few months now, pushing it through some pretty heavy CI/CD pipelines. It's fantastic for the daily grind—but I'm starting to bump into what feels like a performance wall.
My main pipeline now processes deals for about 150 microservices (lots of parallel jobs). I've noticed that once the active deals in a single pipeline run exceed roughly 300-350 individual deal objects, the UI for that pipeline starts to feel noticeably sluggish. We're talking:
* The main pipeline view takes a few seconds to render fully.
* Expanding job groups feels laggy.
* The real-time log streaming is still fine, but navigating the structure is slower.
My setup is pretty standard: we're on the cloud-hosted Granola plan, not self-hosted. I'm curious if this is a known scaling limit based on the frontend rendering that many objects, or if it's more about our particular deal payload size.
Has anyone else pushed Granola to similar scales? Specifically:
* At what number of deals did you start seeing UI or API response degradation?
* Did moving to self-hosted runners (or the enterprise self-hosted Granola option) change that limit significantly?
* Are there any patterns in how you structure your `granola.yml` that help mitigate this? For example, using more, smaller pipelines vs. one giant one?
I love the tool, but for it to be the central hub for our entire deployment process, I need to understand where the practical boundaries are before we design ourselves into a corner.
Ship fast, measure faster.