Repository navigation
Compute GA follow-ups: docs, site, and blog fixes found while reviewing the GA post - #8435
Merged
Merged
Conversation
…ur first app" The porting path imported `postgres` from `@prisma/composer-prisma-cloud`, which the package root does not export (TS2305 with Composer 0.28.0). Use `rawPostgres()`, which delivers `db.url` for the client you already use, and point to `postgres(contract)` from `/orm` for an ORM-typed database. The agent prompt now says the same. The page's `prisma.config.ts` called `prismaCloud()` with no region, so the first deploy of a new project stopped with "no deploy region is configured". Set `region` in the sample, explain `PRISMA_REGION` as the alternative, and name the error in the deploy step. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`new SQL({ url: db.url, max: 1, idleTimeout: 10 })` in the Databases,
Deploying, and "Deploy your first app" pages fails under `npx prisma dev`:
the local Postgres shares one session across client connections, so after
the pool's idle timeout reconnects it, the first query returns
`prepared statement "..." already exists` and the request 500s.
Reproduced with prisma 8.0.0-rc.21 and Composer 0.28.0: with the docs'
options the first request after 15 seconds idle returned 500; with
`prepare: false` added, three rounds of 20 concurrent requests followed
by an idle period all returned 200.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The Compute FAQ, Known limitations, and Request timeout pages said Compute has no cron scheduling and a schedule must come from outside, while the Composer docs ship a `cron` building block that runs a scheduler service on Compute. Readers following one page got the opposite advice from the other. Say what is true on both sides: Compute itself has no built-in cron, and a schedule comes from Composer's `cron` block, which adds a scheduler service that never scales to zero, or from an external scheduler. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… one that doesn't
The Composer overview said "You do not need Composer to deploy to Prisma
Compute" and pointed to Deployments, but every path on that page (deploy
on push, the Console, the CLI's `deploy`) runs a Composer module, so the
two pages read as contradicting each other.
Name the paths that deploy a Composer module (the CLI's `deploy`, the
GitHub deploy action, and the Deploy Button), and add the path that
doesn't to Deployments: the REST API's
`POST /v1/services/{serviceId}/deployments`, which returns an upload URL
for a built artifact and which Alchemy uses. The endpoint is marked
experimental in the OpenAPI spec, and the page says so.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The homepage pricing section, the pricing comparison, and the pricing
calculator said "Unlimited data transfer included", "Data transfer
included", and "No egress fees" in a whole-stack context ("One bill, one
platform"). That holds for Prisma Postgres, which the spec table lists as
unlimited data transfer, but Prisma Compute bills outbound bandwidth at
$0.025 per GB after the Free plan's 10 GB, as the same spec table and the
Compute pricing docs say.
Say "database data transfer" and "database egress" in those places, and
keep the Markdown renditions in step with the components.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
"It works with the coding agents you already use, including Claude, Codex, and Composer" meant Cursor's Composer model. The same launch introduced Prisma Composer, so readers can take the sentence to mean Prisma's deploy tool. Name the product, Cursor, instead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`https://pris.ly/discord-compute` is a 302 to the generic `https://pris.ly/discord` invite, so the link text "the `prisma-compute` channel on our Discord" promised a channel link the URL doesn't deliver. Link the Discord invite directly and keep naming the channel in the text. Pointing the `discord-compute` short link at the channel itself is a change in the pris.ly (Dub) dashboard, outside this repo. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ervice input These pages imported or declared `postgres()` from `@prisma/composer-prisma-cloud`, which the package root does not export (TS2305 with Composer 0.28.0); `postgres` lives in `/orm` and takes a data contract. Their examples want a plain connection, so use `rawPostgres()`, which the root exports and which hands the service `db.url`. The porting agent prompt now names the import too. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The prompt's first step installed only the Composer packages. The `composer` section lives in `prisma.config.ts`, which imports `prisma/config` from the project's own `node_modules`, so an agent that followed the prompt reached `npx prisma dev module.ts` and stopped with `CLI.CONFIG_UNREADABLE`. Install `prisma` as a dev dependency in step 1, as the page's Next.js section already does. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Deploy on push (step 2) and Getting started said that a directory you had deployed from was "already linked", so readers skipped `project link` and then hit `PROJECT.SETUP_REQUIRED` from `git` and `service` commands. In prisma 8.0.0-rc.21 only `project link` and `project create` write `.prisma/local.json`; the Composer `deploy` command never does. Say so, and keep the `project link` step. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The `service` reference said `PRISMA_SERVICE_ID` is the environment fallback for the service argument, and Known limitations told CI to set `PRISMA_PROJECT_ID` / `PRISMA_SERVICE_ID` to override the linked project and service. Neither variable appears anywhere in prisma 8.0.0-rc.21 or @prisma/cli-engine; the only platform id the CLI reads from the environment is `PRISMA_WORKSPACE_ID`. Tell CI to pass `--project` and the service name instead. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…roject
Known limitations said a project created without a region lands in
`us-east-1`. That matches `project create` and the REST API, where the
region is optional, but the first `deploy` of a Composer app stops with
`project "<name>" does not exist yet and no deploy region is configured`
until `prismaCloud({ region })` or `PRISMA_REGION` is set, as the
`deploy` reference says. Add that exception.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
🍈 Lychee Link Check Report362 links: ✅ All links are working!Full Statistics Table
|
Contributor
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configuration
📒 Files selected for processing (21)
Included review availability: This review used your included allowance. 3 included reviews remain after this review. Your included PR review attempts over the past 7 days set your current allowance at 4 reviews per hour. WalkthroughThis pull request updates Prisma Compute and Composer documentation, clarifies database pricing copy across the site, and revises two blog references. The documentation covers database examples, deployment options, project and service targeting, and scheduling. ChangesCompute and Composer documentation
Pricing copy
Blog copy
Priority: ⬇️ Low Estimated code review effort: 3 (Moderate) | ~20 minutes Change: Other Possibly related PRs
Suggested reviewers: Merge Risk: ⚪ Minimal · up to The reviewed documentation and copy changes do not establish a behavior that requires a pre-merge fix. The Composer examples match the package API, so the PR appears ready for normal checks. 🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
Full details: Docstring CoverageExplanation Docstring coverage is 60.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 5 functions across 5 files. (16 skipped: 16 unsupported.)
✨ Finishing Touches 💡 1📝 Generate docstrings 💡
🧪 Generate unit tests (beta)
Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
…c.17 `@prisma/orm-postgres` and `@prisma/orm-mongo` published 8.0.0-rc.17 as `latest` on 2026-10-08, so `pnpm --filter docs lint:versions` fails on every branch, main included, with six stale `8.0.0-rc.16` references. - From scratch, the Solid Start sample output, and the Release status table now show 8.0.0-rc.17 (table dated 8 October 2026). - The PostgreSQL and MongoDB upgrade guides record the version they were written against, so those lines now say "as of 8.0.0-rc.16", which the lint treats as history. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
While reviewing the Prisma Compute GA post (#8433), the review agents found problems in the docs, the site, two blog posts, and two other repositories. This PR fixes the ones in this repository, one commit per item. The rest have their own PRs, linked below.
The items, and where each is fixed
postgresfrom@prisma/composer-prisma-cloud(TS2305: the root doesn't export it) and never sets a region, so a first deploy stops withno deploy region is configuredrawPostgres(), point topostgres(contract)from/orm, setprismaCloud({ region }), explainPRISMA_REGIONSQLsamples (max: 1, idleTimeout: 10) return 500prepared statement "..." already existsunderprisma devafter an idle reconnectprepare: falsein Databases, Deploying, and Deploy your first appcronbuilding blockcronblock (a scheduler service that never scales to zero) or an external schedulerPOST /v1/services/{serviceId}/deployments, used by Alchemy) on Deploymentsprisma-computeagent skill teaches@prisma/cli app deploy,prisma.compute.ts, andcompute:deploy, which no longer existpris.ly/discord-computeis a 302 to the generic Discord invite, so the beta post's "prisma-compute channel" link isn't a channel linkpris.ly/discordand name the channel in the text. Pointing the short link at the channel is a change in the pris.ly (Dub) dashboard, outside any repoAlso fixed here (found by the same reviews)
rawPostgres()instead of the rootpostgres()in Apps and Modules, Porting an existing app, and Service input.prismaas a dev dependency.prisma.config.tsimportsprisma/config, and without itprisma devstops withCLI.CONFIG_UNREADABLE.project linkandproject createwrite.prisma/local.json.PRISMA_SERVICE_IDandPRISMA_PROJECT_IDfallbacks from theservicereference and Known limitations. Neither variable exists in prisma 8.0.0-rc.21.deployneeds a region for a new project.How it was checked
tscon the corrected samples, the CLI source for linking and environment variables, and--help.prepare: falsefix was reproduced undernpx prisma dev. Before it, the first request after an idle period returned 500. After it, three rounds of 20 concurrent requests, each followed by an idle period, all returned 200.pnpm lint:linksin apps/docs and apps/blog: 0 errors.check-ai-signs.sh(whatdocs-proseruns) passes on every changed page.Found but not fixed here
The skills rewrite turned up more places where the docs and the CLI disagree. The CLI wins in each case:
agentcommand group, which returnsCLI.UNKNOWN_COMMAND.FEATURE_UNAVAILABLE; the CLI returnsSERVICE.LOGS_FAILED/SERVICE.LOGS_INCOMPLETE.--branch, otherwise the production branch.deployreference: the control-APIdeploy/destroysnippets omitconfig, which is required in 0.28.0 (tscfails).rmaliasesremove. The command isdelete, with no aliases.service domain remove, and Domains and Deployments say--yesskips confirmation. The command isdelete, and it needs--confirm <hostname>.ver_…in the docs,cpv_…in CLI help.deploybuilds and shows it without an entry argument.Unrelated CI fix included
@prisma/orm-postgresand@prisma/orm-mongopublished8.0.0-rc.17aslateston 2026-10-08, which brokepnpm --filter docs lint:versions("Validate documentation links") on every branch, main included. One separate commit bumps the six stale8.0.0-rc.16references: the current-version mentions now sayrc.17, and the two upgrade guides mark the version they were written against with "as of". The same commit is on #8433 and #8435, so whichever merges first carries it.🤖 Generated with Claude Code
Summary by CodeRabbit