I enabled enhanced observability and found some great insights from it, the feature is really good. However I had to disable it permanently because of the complete lack of control over how the events are billed
There is no way to disable enhanced observability on a per project basis. I have one project that I want observability on but most of my quota is being used by another project that I don’t really want enhanced observability on.
That project where I want to disable enhanced observability is a frequent target for aggressive scrapers and AI bots. I have quite aggressive Vercel WAF rules to control my platform usage and block these bots. However, blocking a request with Vercel WAF does not prevent billing for enhanced observability. This makes the feature super risky, as even though I am doing everything in my power to protect myself I still can’t control the usage on this feature.
I find this behaviour very poor from Vercel, it breaks my expectations of how WAF is supposed to work. WAF isn’t fully protecting me. I’m also literally paying for data that I can’t access. This behaviour was confirmed by support as “how the monitoring system works” and not a bug, super frustrating.
So, three requests:
- Vercel WAF blocks shouldn’t incur observability event usage
- In the meantime, add warnings into the Vercel docs to call out this risk
- Allow disabling enhanced observability on a project-by-project basis