Error calling enhanced Prisma methods when using Prisma Accelerate

[gczh]

Description and expected behavior
I was setting up Prisma Accelerate to reduce latency and mitigate cold start times of my Vercel functions.

My prisma is enhanced through zenstack. When using enhancedPrisma for queries, I get errors in the screenshot below.

I've also tried doing npx zenstack generate --no-engine and it doesn't work

Screenshots

Environment (please complete the following information):

• ZenStack version: 1.7.0 (and "@zenstackhq/runtime": "1.7.0")
• Prisma version: 5.7.1
• Database type: Postgresql

Additional context
@ymc9 mentioned on Discord the following:

It seems to be caused by the default large transaction timeout used internally by ZenStack. Looks like a new limit set by Prisma Accelerate.

In V2 we're adding an option to configure it. Maybe we should integrate that change to the next V1 release to unblock.

[ymc9]

Hi @gczh , I've pushed a 1.8.2 build live. Tested with Prisma 5.9.1 with accelerate and I haven't found any other problem so far.

What you need to change is to pass in the two transaction options and use smaller values, like:

enhance(
  prisma,
  { ... },
  {
    transactionTimeout: 10000,
    transactionMaxWait: 10000,
  }
);

I've also added a new option to the "@Core/prisma" plugin, which you can use to append additional arguments to the prisma generate command run automatically by zenstack generate:

plugin prisma {
  provider = "@core/prisma"
  generateArgs = "--no-engine"
}

With that, you don't need to explicitly run prisma generate --no-engine. https://zenstack.dev/docs/reference/plugins/prisma
Please let me know if it works for you

[gczh]

@ymc9 Tested it and it works great! Huge huge thanks to ya for pushing the updates 🙌