Hey everyone! 👋 I'm diving into setting up error tracking for our SaaS app and I've hit a bit of a crossroads. We handle a pretty high volume of user events, and our current logging is... well, let's just say it's not cutting it when we need to find the root cause of a bug fast.
I've narrowed it down to Sentry and Rollbar, since they both seem popular in the Python/Flask world we're in. But I'm a bit stuck on the real-world differences when you're dealing with thousands of errors per minute. I get the basic feature lists from their sites, but I'm hoping for some practical insights.
For those who've used either (or both!) in a high-volume environment:
* How do they really compare on ingestion costs when you're dealing with millions of events? Is the pricing model straightforward or does it get tricky with spikes?
* We have some pretty high-cardinality data (lots of unique user IDs, session tags). Does one handle that dimensionality better without melting down or becoming insanely expensive?
* What's the query latency like when you're trying to filter and search through a huge backlog of errors during an incident?
* And maybe most importantly, how reliable are the alerts? Do they ever get drowned out or delayed during a spike?
I'm leaning towards trying to orchestrate this with Airflow eventually, so any integration quirks would be good to know too. I just want to make sure I'm picking a tool that scales with us and doesn't become a bottleneck itself.
Thanks in advance for saving this rookie from a potential architectural misstep!
-- rookie
rookie