## Problem
The Phase 1 mw_diag monitoring dashboard wasn't showing any data despite
the middleware PR (#43413) being live in production.
Next.js `response.cookies.set()` URL-encodes cookie values, so
`hit=1&would_stamp=0&has_cookie=0` becomes
`hit%3D1%26would_stamp%3D0%26has_cookie%3D0` in `document.cookie`. Our
`parseMwDiagCookie` was passing that encoded string directly to
`URLSearchParams`, which splits on literal `&` and `=` — the encoded
`%26`/`%3D` delimiters were invisible to it. Result: `params.get('hit')`
always returned `null`, so every diagnostic cookie was silently dropped.
## Changes
- Add `decodeURIComponent()` before `URLSearchParams` in
`parseMwDiagCookie`
- Add 3 test cases covering URL-encoded cookie values (matching what
Next.js actually produces in production)
## Testing
- All 34 unit tests pass (31 existing + 3 new)
- Verified locally with Playwright: manually set the URL-encoded cookie
format on localhost:8082, confirmed `parseMwDiagCookie` now returns
`{hit: true, ...}` and the data flows through to `PageTelemetry`
GROWTH-625
When saving Data API settings with only max rows changes (no
table/function privilege changes), updateExposedEntities sent an empty
string to pg-meta/query which failed server-side validation with 'String
must contain at least 1 character(s)'. Early return when there are no
SQL parts to execute.
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
Docs update
## What is the new behavior?
Adds visual and more context to the egress troubleshooting guide. Adds
major updates to edge function's 546 troubleshooting guide
---------
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
docs update
## What is the current behavior?
n/a
## What is the new behavior?
n/a
## Additional context
Add any other context or screenshots.
---------
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
doc
## What is the current behavior?
Doc incorrectly states physical backups are not enabled by default.
## What is the new behavior?
Backup doc page now correctly clarifies physical backups are enabled by
default,
assuming project is eligible.
## Additional context
NA
---------
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
This PR changes the logic for show/hide password icon in the sign up
form. Fix#39399 and #41459
## What is the current behavior?
For Password Field Eye Icon Label
<ul>
<li>When password is hidden: Button shows "Hide password" label
(incorrect).
<li>When password is visible: Button shows "Show password" label
(incorrect)
</ul>
For Edge Functions Secrets
<ul>
<li>All secret values are always visible in plain text
<li>No visibility toggle functionality exists
<li>No way to hide/show individual secret values
</ul>
## What is the new behavior?
Sign-up Form Password Field:
<ul>
<li>When password is hidden: Button correctly shows "Show password"
label
<li>When password is visible: Button correctly shows "Hide password"
label</ul>
Labels now accurately describe what clicking the button will do
<p></p>
Edge Functions Secrets:
<ul>
<li>Secret values are hidden by default (security-first approach)
<li>Each secret has its own independent visibility toggle button
<li>Hidden secrets display as bullet points (•••••••••••••••••••••)
<li>Eye icon appears when secret is hidden - click to reveal
<li>EyeOff icon appears when secret is visible - click to hide
<li>Tooltip shows "Hidden" when value is masked, actual value when
visible</ul>
---------
Co-authored-by: Gildas Garcia <1122076+djhi@users.noreply.github.com>
Co-authored-by: Alaister Young <a@alaisteryoung.com>
Remove duplicate table assignment using undefined `supabase` variable
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
docs update
## What is the current behavior?
-
## What is the new behavior?
-
## Additional context
-
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Documentation**
* Simplified the Python API client examples in the generating Python
types guide by removing redundant code assignments, making the
documentation clearer and easier to follow.
<sub>✏️ Tip: You can customize this high-level summary in your review
settings.</sub>
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
docs update, added my name
## What is the current behavior?
N/A
## What is the new behavior?
N/A
## Additional context
Onboarding task
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Chores**
* Updated team roster documentation.
<sub>✏️ Tip: You can customize this high-level summary in your review
settings.</sub>
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
**Changes:**
- Add function exposure controls to Data API settings
- Adds a per-function GRANT/REVOKE UI to the Data API settings page,
alongside the existing table selector.
- Empty state: shows "No tables/functions available" instead of "0 of 0
X exposed" when there's nothing to show
- Renamed update-exposed-tables-mutation →
update-exposed-entities-mutation to handle both tables and functions in
one save
**Decisions of note:**
- Functions are grouped by schema.name across overloads — toggling one
entry grants/revokes `EXECUTE` on all overloads
- Only normal and window functions are shown (`prokind in ('f', 'w')`) —
triggers and aggregates are excluded
---
**To test:**
- Table and schema selectors still works end-to-end
- Toggling a function and saving updates the count label and applies the
grant (check via SQL Editor: select * from
information_schema.role_routine_grants where routine_schema = 'public')
- Functions in unexposed schemas are dimmed and unclickable
- With no functions in the selected schemas, the selector label reads
"No functions available" and the dropdown shows the same
## Problem
When editing email templates or edge functions, users may navigate away
from the page without a warning indicating they may loose their changes.
This is because we only handle the `beforeunload` event when we should
also handle NextJS routing events.
This is actually done for the observability reports.
## Solution
Extract the logic from the observability reports into a reusable
component and use it where needed
## How to test
On staging, for each case:
- Authentication email templates
- Observability reports
- Edge functions creation
- Edge functions edition
Test:
- Modify the template/report/function
- Navigate away using either the sidebar link, the browser back button
or closing the tab
- Cancel navigation in the confirmation dialog
- Navigation should be prevented and you should not loose your changes
Test:
- Modify the template/report/function
- Navigate away using either the sidebar link, the browser back button
or closing the tab
- Confirm navigation in the confirmation dialog
- Navigation should not be prevented and you should have lost your
changes
Updated the secret creation command for Supabase Vault to use 'anon_key'
instead of 'publishable_key'.
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
docs
https://supabase.com/docs/guides/functions/schedule-functions
<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit
* **Documentation**
* Updated code examples in the schedule functions guide to reflect
correct Supabase key naming conventions.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
Co-authored-by: Chris Chinchilla <chris.ward@supabase.io>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
Feature — A/B experiment for a new pricing calculator on the `/pricing`
page, gated behind a PostHog feature flag.
## What is the current behavior?
The pricing page shows a single compute pricing section
(`PricingComputeSection`) to all visitors with no experimentation
support.
## What is the new behavior?
- Introduces a **pricing calculator A/B experiment**
(`pricingCalculatorExperiment`) using PostHog feature flags evaluated
client-side.
- **Control variant**: Renders the existing `PricingComputeSection` (no
change).
- **Test variant**: Renders a new `NewPricingComputeSection` with:
- A plan selector toggle (Pro / Team)
- An interactive compute calculator with slider-based instance sizing
- Ability to add/remove multiple project instances
- Live monthly cost estimate breakdown (plan + compute - credits)
- Expandable compute pricing table with detailed specs
- Adds `getFeatureFlag` and `onFeatureFlags` methods to the shared
`PostHogClient` for client-side flag evaluation on www pages (where
server-side evaluation lacks full person context).
## Additional context
- The experiment is scoped to the `/pricing` page only.
- New components (`NewPricingComputeSection`,
`NewComputePricingCalculator`) are created alongside the existing ones —
the control path is completely untouched.
- Feature flag evaluation happens client-side because www pages don't
have full person context on the server.
- Telemetry type `PricingCalculatorExperimentExposedEvent` is added to
`telemetry-constants.ts`.
---------
Co-authored-by: Sean Oliver <882952+seanoliver@users.noreply.github.com>
Co-authored-by: Mert YEREKAPAN <mertyerekapan@gmail.com>
Co-authored-by: Mert YEREKAPAN <33198490+myerekapan@users.noreply.github.com>
## Summary
The `homeNew` PostHog experiment has concluded. This PR graduates it by
making the new homepage (`ProjectHome`, formerly `HomeV2`) the permanent
default for all users, and removes all dead code from the old
experiment.
## Changes
- Remove `homeNew` PostHog feature flag checks and `home_new` experiment
exposure tracking from 3 files
- Rename `HomeNew/` → `ProjectHome/` directory and `HomeV2` →
`ProjectHome` export
- Delete old `Home/Home.tsx` component (shared components like
`ProjectList/` are kept — still used by org pages)
- Delete `pages/project/[ref]/building.tsx` and add a server-side
redirect from `/project/:ref/building` → `/project/:ref` to prevent 404s
during rollout (old cached JS bundles may still route to `/building`)
- Simplify `ContentWrapper` building-state logic in `ProjectLayout` —
always redirect building projects to home, always suppress building
interstitial on home page
- Always route to `/project/{ref}` after project creation (remove
`/building` path)
- Update all Observability imports from `HomeNew` → `ProjectHome`
## Self-hosted behavior change
Self-hosted Studio previously showed the old `Home` component (client
libraries + example projects) since PostHog flags don't load. This PR
changes self-hosted to show `ProjectHome` (TopSection with service
status + instance diagram, advisor, custom reports). All sections query
backend APIs that exist on self-hosted. E2E tests pass against the
self-hosted build.
## Testing
- [x] `pnpm turbo run build --filter=studio` passes
- [x] No remaining references to `homeNew`, `home_new`, or `HomeNew` in
codebase
- [x] No broken imports to deleted files
- [x] Self-hosted E2E tests pass (145 passed, 1 flaky, 4 skipped)
- [x] `/building` redirect added to both platform and self-hosted config
blocks
**Quick test:**
1. Navigate to any project homepage — should render the ProjectHome
component
2. Create a new project — should redirect to `/project/{ref}` (not
`/building`)
3. Visit a project in `COMING_UP` state on a non-home route — should
redirect to home
4. Visit `/project/{ref}/building` directly — should 302 redirect to
`/project/{ref}`
## Linear
- fixes GROWTH-671
## Context
We reached to a consensus that we'll keep manually written queries
within the `pg-meta` package
I'll shift queries over in several PRs but wanted to double check if the
changes here is a good folder structure
## Changes involved
- Am organizing dashboard queries in `pg-meta` based on product
- So folder structure will be `/sql/studio/[product]/*.ts`
- Every constant will be exported directly from `pg-meta`'s index
- So consumption in dashboard will just be `import { xxx } from
'@supabase/pg-meta`
cc @avallete and @soedirgo on this one - lemme know if this is alright,
or if there's a preferred way to do this
PostHog flagged our SDK as outdated — we were on 1.257.2, latest is
1.357.0.
**Changes**
- Bumped `posthog-js` in `packages/common/package.json` from `^1.257.2`
to `^1.333.0` (resolved to 1.357.0)
- Updated lockfile
**Why is the lockfile diff so large?**
posthog-js added several new dependencies between 1.257.2 and 1.357.0: a
full `@opentelemetry/*` stack (for opt-in log exporting),
`@posthog/core` and `@posthog/types` (internal monorepo split),
`dompurify`, and `query-selector-shadow-dom`. None of these affect our
bundle unless explicitly enabled — they're opt-in features. The lockfile
entries are large because pnpm records each package with its full
resolution tree and checksums.
**Testing**
Spun up the full local stack and verified all three PostHog endpoints
return 200 with the new SDK version:
- `/e/` (event capture) — pageviews firing correctly
- `/flags/` (feature flag evaluation) — flags loading correctly
- `/i/v0/e/` (identify) — user identification working
No breaking changes in any of the 16 releases between our old and new
version (all Minor/Patch Changes). TypeScript compilation clean across
all apps.
GROWTH-589
Bug fix
## What is the current behavior?
Read-only users cannot download files because the download feature
requires minting a temporary API key, which is properly blocked for
read-only users.
## What is the new behavior?
Instead of using temporary API keys, we now create signed URLs for the
files to be downloaded. We batch-create signed URLs for an entire
folder's worth of files, requiring only a single management API call,
then use those signed URLs to download the files. This allows read-only
users to download files without needing elevated permissions.
## Additional context
Resolves FE-2737
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
- Added filter by value for the new feature bar experience
- Will follow up with an E2E test later
## Problem
The `_sb_first_referrer` cookie isn't working. The www middleware
matcher explicitly excludes `/dashboard` and `/docs`, so the cookie
never gets stamped for Studio or Docs traffic. PostHog confirmed: only 1
event with `first_referrer_cookie_present=true` out of ~46.5M Studio
pageviews in the last 7 days.
## Background: what the matcher does
In Next.js, the `matcher` config controls which incoming requests the
middleware function even runs on. If a path doesn't match, the
middleware is skipped entirely — the request passes through untouched.
If it matches, the middleware runs and can mutate the response (set
cookies, headers, etc.).
This matters because Studio's SPA navigation works via silent
`/_next/data/` JSON fetches. If middleware runs on those requests and
returns `NextResponse.next()` with any mutations, it breaks those
fetches and causes full page reloads instead of client-side transitions.
## What we tried before
| PR | www runs on `/dashboard`? | Studio `proxy.ts` runs on all routes?
| Result |
|---|---|---|---|
| **#42768** (Attempt 1) | ✅ Yes — and also intercepts `_next/data` | ✅
Yes — `matcher` config removed, stamps cookie everywhere | Full page
reloads in Studio |
| **#43129** (Full revert) | ❌ No — www middleware deleted entirely | ❌
No — restored to `matcher: '/api/*'` only | Back to baseline, no cookie
stamping anywhere |
| **#43153** (Attempt 2) | ❌ No — `/dashboard` explicitly excluded | ✅
Yes — `matcher` config removed again, stamps cookie everywhere | Full
page reloads in Studio again |
| **#43189** (Attempt 3) | ❌ No — same as #43153 | ✅ Yes — `matcher`
config still removed, cookie stamping made conditional | Still broken |
| **#43190** (Ivan's fix) | ❌ No — `/dashboard` still excluded | ❌ No —
restored to `matcher: '/api/*'` only | Works — but cookie never stamps
for `/dashboard` traffic |
| **#43413** (this PR) | ✅ Yes — sets diagnostic cookie only, no
attribution stamping yet | ❌ No — unchanged, still `matcher: '/api/*'`
only | ❓ Untested in prod |
The common factor in every failure: Studio's `proxy.ts` ran on all
routes (including `_next/data` requests), which broke SPA navigation.
This PR is the first one that runs www middleware on `/dashboard` while
Studio's `proxy.ts` stays in its original narrow `/api/*` scope.
## What changed
This is Phase 1 of a two-phase rollout. We remove `dashboard|docs` from
the matcher's negative lookahead so www middleware runs on those paths —
but instead of stamping cookies, we set a short-lived (60s) diagnostic
cookie `_sb_mw_diag` on `/dashboard` and `/docs` requests.
The diagnostic cookie encodes
`hit=1&would_stamp={0|1}&has_cookie={0|1}`, which Studio telemetry reads
on the initial pageview and reports to PostHog as `mw_diag_hit`,
`mw_diag_would_stamp`, and `mw_diag_has_existing_cookie` properties.
A cookie rather than a header because response headers aren't readable
by JS. It also tests the actual Set-Cookie mutation path that Phase 2
will use (which is what Next.js issue #41885 is specifically about).
Phase 1 answers two key questions before we commit to Phase 2:
1. Does expanding the matcher break Studio SPA navigation?
2. What % of /dashboard arrivals would get a first-referrer cookie
stamped in Phase 2?
## Phase 2 readiness criteria
**Important caveat**: `mw_diag_*` data reflects consented users only and
may under-represent first-visit anonymous traffic. The Phase 2 decision
should account for this — the actual middleware execution rate is likely
higher than what PostHog reports.
### PostHog query spec
**Middleware execution rate**: Of all Studio initial pageviews on
`/dashboard` or `/docs` paths, what percentage have `mw_diag_hit =
true`? Expected: >= 90%. Below 70% warrants investigation (could
indicate edge caching bypassing middleware, or a matcher configuration
issue).
```
Filter: event = "$pageview" AND (current_url contains "/dashboard" OR current_url contains "/docs")
Breakdown: mw_diag_hit (true vs null/missing)
Metric: count(mw_diag_hit = true) / count(all) * 100
```
**Would-stamp rate**: Of events with `mw_diag_hit = true`, what
percentage have `mw_diag_would_stamp = true`? This tells us what
percentage of Phase 2 traffic would actually get a cookie stamped. No
hard threshold — unexpected values (< 5% or > 95%) suggest a logic bug
worth investigating before Phase 2.
```
Filter: event = "$pageview" AND mw_diag_hit = true
Breakdown: mw_diag_would_stamp (true vs false)
Metric: count(mw_diag_would_stamp = true) / count(all) * 100
```
**Existing cookie rate**: Of events with `mw_diag_hit = true`, what
percentage have `mw_diag_has_existing_cookie = true`? This tells us how
many users already have the cookie from a prior www visit.
```
Filter: event = "$pageview" AND mw_diag_hit = true
Breakdown: mw_diag_has_existing_cookie (true vs false)
Metric: count(mw_diag_has_existing_cookie = true) / count(all) * 100
```
### Go / no-go threshold table
| Signal | Go | Investigate | No-Go |
|---|---|---|---|
| `mw_diag_hit` rate (% of /dashboard+/docs pageviews) | >= 90% | 70-90%
| < 70% |
| SPA navigation errors (Sentry / Vercel logs) | No increase | < 0.1%
increase | > 0.5% increase |
| Middleware p99 latency (Vercel function logs) | < 50ms added |
50-100ms | > 100ms |
| Sample volume in first 24h | > 1,000 events | 100-1,000 (extend
window) | < 100 (insufficient data) |
## Changes
- `apps/www/middleware.ts`: Removed `dashboard|docs` from matcher; added
`isDashboardOrDocs` guard that sets `_sb_mw_diag` diagnostic cookie
instead of stamping attribution
- `apps/www/middleware.test.ts`: 12 tests covering cookie stamping on
www paths, diagnostic cookie encoding for all scenarios (external
referrer, direct nav, internal referrer, existing cookie)
- `packages/common/first-referrer-cookie.ts`: Exported
`MW_DIAG_COOKIE_NAME`, `MwDiagData` type, and `parseMwDiagCookie()`
helper
- `packages/common/telemetry.tsx`: Reads `_sb_mw_diag` on initial Studio
pageview; reports `mw_diag_*` properties to PostHog
## Testing
Unit tests pass (13/13 www, 31/31 first-referrer-cookie). Production
validation needed:
- [ ] Studio SPA navigation works (tab changes, SQL editor, no full page
reloads)
- [ ] PostHog shows `mw_diag_hit = true` on Studio initial pageviews
- [ ] `mw_diag_would_stamp` distribution looks reasonable before
enabling Phase 2
- [ ] Monitor 24h before Phase 2
Ref: GROWTH-625 / GROWTH-668
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
This is something we've discussed for a while. A shared place for error
codes and their mappings, descriptors etc. Available for all projects in
the monorepo.
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
This introduces a small snippet preview of the error code coming via our
content API for the docs. This appears in a couple of places right now.
- **Auth Overview** - This page isn't fully released yet, but it appears
on the error codes table (as depicted below).
- **Logs** - When you delve into the logs panel, if there's an error
code available, they're also wrapped in this popover.
We are also working on a shared-data package (#43458) to potentially
replace this endpoint internally. Also introduces a multifaceted button
to debug/fix with either Assistant or LLM of choice.
| Auth Overview | Logs Panel |
|--------|--------|
| <img width="407" height="287" alt="Screenshot 2026-03-09 at 14 14 09"
src="https://github.com/user-attachments/assets/7450dddb-6828-4cd3-802d-37d47ba1b440"
/> | <img width="394" height="216" alt="Screenshot 2026-03-09 at 14 13
56"
src="https://github.com/user-attachments/assets/80c2a46e-dbe4-4e88-a0a7-68b977a71d6b"
/> |
---------
Co-authored-by: Claude Sonnet 4.6 <noreply@anthropic.com>
Lays groundwork for online evals on Assistant chat logs.
https://www.braintrust.dev/docs/observe/score-online
### Changes
- New workflows:
- `braintrust-scorers-deploy.yml` keeps prod scorers in sync on push to
`master`
- `braintrust-preview-scorers-deploy.yml` deploys preview scorers to the
staging project for PRs labeled `preview-scorers`, posting a comment
with scorer links
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000097222))
- `braintrust-preview-scorers-cleanup.yml` deletes preview scorers when
the PR is closed
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000749847))
- Adds `evals/scorer-online.ts` entry point invoked with `pnpm
scorers:deploy`, registering scorers for online evals in the Braintrust
"Assistant" project
- Refactors scorer code to separate online-compatible scorers
(`scorer-online.ts`) from WASM-dependent ones (`scorer-wasm.ts`)
- "URL Validity" scorer now only checks Supabase domains to prevent
requests to untrusted origins
- Span `input` is now shaped `{ prompt: string }` instead of plain
`string` for compatibility with offline eval scorers
- Env vars `BRAINTRUST_STAGING_PROJECT_ID` and `BRAINTRUST_PROJECT_ID`
configured in GitHub repo settings
- `generateAssistantResponse` now uses `startSpan` + `withCurrent`
instead of `traced()` to manually manage the root span lifecycle — this
ensures `onFinish` logs output to the span _before_ `span.end()` is
called, which is when Braintrust triggers scoring automations
### Online Scorers
We share scoring logic across offline and online evals, but some of our
scorers aren't transferrable to an "online" setting due to runtime
challenges or ground truth requirements.
**Supported**
- Goal Completion
- Conciseness
- Completeness
- Docs Faithfulness
- URL Validity
**Unsupported**
- Correctness (requires ground truth output)
- Tool Usage (requires ground truth requiredTools)
- SQL Syntax (uses libpg-query WASM)
- SQL Identifier Quoting (uses libpg-query WASM)
### How to use these scorers
Going forward if you want to add/edit online eval scorers, add the
`preview-scorers` label to a PR. This deploys scorers to the [Assistant
(Staging
Scorers)](https://www.braintrust.dev/app/supabase.io/p/Assistant%20(Staging%20Scorers)?v=Overview)
project in Braintrust with branch-specific slugs, and comments on the PR
([example](https://github.com/supabase/supabase/pull/43194#issuecomment-4000097222)).
From the Braintrust dashboard you can "Test" the scorer with traces from
any project.
<img width="1866" height="528" alt="CleanShot 2026-03-05 at 15 15 00@2x"
src="https://github.com/user-attachments/assets/4f15cebc-3f2d-4e8a-9ee2-fe8ef7bf4199"
/>
Once merged, scorers are deployed to the primary
[Assistant](https://www.braintrust.dev/app/supabase.io/p/Assistant)
project, and preview scorers are deleted from the staging project. Down
the road, scorers on the Assistant project will run automatically on a
sample of production traces.
Closes AI-437
## Problem
- database e2e tests run in serial mode, which is slower
- they also are a bit flaky
## Solution
- Ensure they can run in parallel
- Make them more stable by using UI checks when waiting
- Use assertions that automatically wait/retry in playwright
Follow up of #43547 and #43560
## Problem
The cron job tests run in serial mode and can be flaky.
## Solution
- Use simple query to install the `pg_cron` extension (way faster than
using UI)
- Make sure all tests can run in parallel
## Problem
The `tableCreateGeneratePolicies` A/B experiment tested an AI-assisted
policy generator (`RLSManagement`) as a variation in the table creation
side panel. The variation didn't pass — control wins, so the standard
RLS checkbox should be the permanent behavior.
## Changes
- Deleted `useTableCreateGeneratePolicies` hook (PostHog flag + exposure
tracking)
- Deleted the entire `RLSManagement/` component directory (variation UI
— `RLSManagement`, `PolicyList`, `PolicyListEmptyState`,
`ToggleRLSButton`)
- Removed experiment flag checks from `TableEditor.tsx` — RLS checkbox
is now always rendered (was already the control)
- Removed experiment conversion tracking from `SidePanelEditor.tsx`
## Testing
Verified `pnpm typecheck` passes clean. Table creation side panel
renders the RLS checkbox unconditionally as it did in the control.
GROWTH-653
## Problem
The `projectCreationEnableRlsEventTrigger` A/B experiment tested showing
an opt-in "Enable automatic RLS" checkbox in the project creation flow,
which sets up a Postgres event trigger to auto-enable RLS on every new
table in the public schema. The test variant passed with a +3.3pp lift.
## Changes
- `SecurityOptions.tsx` — removed PostHog flag check, checkbox is now
always rendered
- `pages/new/[slug].tsx` — removed experiment flag reads, exposure
tracking (`useTrackExperimentExposure`), and the conditional
`rlsOptionVariant` telemetry property
The underlying feature logic (`enableRlsEventTrigger` form field,
`AUTO_ENABLE_RLS_EVENT_TRIGGER_SQL`, submission handling) is unchanged —
we're just removing the scaffolding that was gating it.
## Testing
Verified `pnpm typecheck` passes clean. The "Enable automatic RLS"
checkbox now shows unconditionally in the Security Options section of
project creation.
GROWTH-653
## What kind of change does this PR introduce?
Remove feature flag.
## What is the current behavior?
The [State of Startups](https://supabase.com/state-of-startups) page
instantly redirects to the homepage rather than staying put.
## What is the new behavior?
The [State of Startups](https://supabase.com/state-of-startups) page
rightfully loads.
## Additional context
Some sort of race condition with the `useFlag`. Simpler to just remove
the flag logic entirely and keep the page permanently on.
## Problem
- Some tests don't always cleanup after themselves or don't do it
consistently which make it hard to work on them locally
- Some test suites don't allow parallel execution of their tests
- Some tests are flaky
This a follow up of #43547
## Problem
e2e tests are still flaky and not as fast as they could be
## Solution
- [x] Reset supabase instance: this makes the database visualiser tests
faster as there are less tables to screenshot
- [x] Improve the single file setup utilities so that they never block
local tests by cleaning them up before starting
- [x] Disable animations while running the tests (less time waiting for
animations to complete
- [x] Add utility functions that help reproducing flaky tests locally
This PR updates @supabase/*-js libraries to version 2.99.0.
**Source**: supabase-js-stable-release
**Changes**:
- Updated @supabase/supabase-js to 2.99.0
- Updated @supabase/auth-js to 2.99.0
- Updated @supabase/realtime-js to 2.99.0
- Updated @supabase/postgest-js to 2.99.0
- Refreshed pnpm-lock.yaml
This PR was created automatically.
Co-authored-by: supabase-releaser[bot] <223506987+supabase-releaser[bot]@users.noreply.github.com>
## I have read the
[CONTRIBUTING.md](https://github.com/supabase/supabase/blob/master/CONTRIBUTING.md)
file.
YES
## What kind of change does this PR introduce?
doc update
## What is the current behavior?
Error description states that realtime breaks when a message is JSON
parsable. Should state when it is "not" JSON parsable
## What is the new behavior?
Added not key word to description:
```
Payload sent in NOTIFY operation was "NOT" JSON parsable.
```
Sean Oliver