Understanding compute time pricing rate

Hi team,

First of all, really cool product, and enjoying the dev experience.

I was a bit confused about my bill this month. I havent used Neon in any active projects yet, so all compute has been from me during development. So I was surprised that this amounted to $0.56 for 4.72 compute units.

I suppose the compute units count all hot time with a minimum %cpu. So a single user making a request every minute is relatively expensive, but for 100 light users the price for compute wouldnt go up much. Is that a fair assumption?

What surprised me was that rate on the invoice was 0.118 for europe, but on the website the rate is listed at 0.0295. What am I missing here?


Let me close this question, because I think I figured it out already.

I think I get that the 4.72 compute units means 4*4.72 times the minimum compute (0.25), which is priced at 0.0295. So what supposedly happened is that during development I ran up about 18 hours of active time. So I was about as expensive as a single user could get by repeatedly triggering the 5 min active time.

Not that 0.56 is expensive, off course, but I needed to be sure that this was far from a normal estimate for a single user.