Hi everyone,
I've been evaluating Langfuse for our team's LLM observability needs, and overall, I'm impressed with the core feature set. However, I wanted to flag something our team ran into during the pricing assessment that wasn't immediately obvious from the public pricing page.
We operate with standard dev/staging/production environments. The Langfuse model of pricing per "project" seemed straightforward at first. But when we dug into setting it up, we realized that for a clean, isolated multi-environment setup, you essentially need to create a separate "project" for each environment. Since pricing is per project, this means the monthly cost scales linearly with each environment we spin up.
This wasn't a deal-breaker for us, but it was a significant adjustment to our initial cost estimate. It feels a bit like a hidden multiplier. I'm wondering how others are handling this.
* Are you creating separate projects for dev/staging/prod?
* Are you instead using a single project and using tags/filters to separate environments? If so, how's that working for data isolation and team permissions?
* Have you found any clever workarounds or has the Langfuse team offered any guidance on best practices here?
For a tool that's so developer-focused, I'd expect the pricing to more naturally accommodate a standard multi-environment workflow without this surprise scaling. Keen to hear how the community is navigating this.