Hey everyone 👋 a question around Prisma Data Platform and Planetscale: What's the advantage of setting up Proxy in front of Planetscale? If I understand correctly it's a serverless db that is supposed to scale connections on its own. Is it just so we can use the Data Browser and reduce the size of the lambdas? Or are there other things I am not aware of? (Reason I'm worrying: Planescale region is in Dublin, Prisma Proxy is in Frankfurt, maybe it adds unnecessary delays)

Hello Francesco, I think short answer is “it depends”. If you are using PS from serverless, you likely to have a large number of functions open concurrent connections to your database. The difference with PlanetScale is that it supports a higher number of these than other DBs. AFAIK, the current limit is 1K in the free Developer plan and 10K in the Scaler plan. Other DB providers have often lower limits for these so it’s much easier to exhaust the connection limit. If this number is enough to satisfy your traffic workload then you are good from the connection limit POV. The one thing that you might still experience is cold starts. Serverless functions are short-lived ephemeral and rarely get reused. This means that more often than not a request will be served by a newly spawned function, which needs to first open a persistent connection to the database. This has some latency cost associated with it, that will depend on the database and provider, e.g. AFAIK Postgres connections take longer to establish than MySQL. We plan to offer support for new regions soon and also offer additional functionalities down the road. Would you use the Data Proy better if you had the ability to deploy it in Dublin? If so, please request the region in this form here.

Hey thank you Alberto that is a very comprehensive answer. I don't know I'm just tinkering with it at the moment so the Dublin <> Frankfurt trip is not an issue for me but I was wondering if it was something worth considering. The explanation regarding open connections and cold start is very compelling so I might try it as it is and see if I notice any considerable slowdowns. The single biggest feature that drives me towards Prisma Data Platform is the "Preview Database" that was hinted in the last conference. Do you have any insight on when it will be open as a beta to try it ouy?

Hello Francesco, I don’t of timeline for this. My colleague @Spiros - Product at Prisma might be able to share more information.

Hello @Francesco Sardo! We are looking at a timeline beyond Q1 '22. It's currently on top of our backlog ("Candidates" column in our public roadmap), so once it's picked up you should see it move to the right. I will also, most probably, reach out to you to test some implementation details of the feature with you 🙂

Hey @Spiros - Product at Prisma , looks great I would be happy to

@Spiros - Product at Prisma I'd love to be added to that list of people willing to test the implementation details!