Query insights gain visibility into sharded Vitess workloads
June 29 – July 5, 2026
PlanetScale brought tag-based filtering and vindex usage tracking to Query Insights for Vitess databases, while Neon released a new TypeScript SDK and Supabase shipped a major CLI update with pipeline-incompatible migration support.
Deeper observability for sharded databases
PlanetScale delivered two significant enhancements to Query Insights for Vitess databases this week. Tag-based filtering allows developers to classify queries by application, route, or job using built-in metadata or SQLCommenter tags, then filter the query table with tag:key:value syntax. The new Tags view surfaces breakdowns, and individual query executions show their attached tags. Complementing this, vindex usage tracking now displays the percentage of traffic that used relevant vindexes alongside a time-series graph, making it straightforward to spot queries missing indexes and measure the impact of index changes.
PlanetScale also added VTTablet and MySQL configuration directly from the Clusters page and API, and the PlanetScale MCP server gained replica targeting for read queries (with an opt-out to route to the primary) plus automatic detection of Postgres RLS policies that might filter query results. The Terraform provider v1.3.0 introduced planetscale_postgres_redacted_branch_role, a Postgres resource that keeps passwords out of Terraform state entirely.
New SDKs and CLI workflows
Neon's biggest story was the launch of @neon/sdk, a fetch-based, zero-dependency TypeScript client for the Neon API that replaces @neondatabase/api-client. The SDK is ESM-only and generated from the OpenAPI spec, with an ergonomic layer handling authentication, typed errors, automatic retries, readiness polling, and multi-step workflows like snapshot preview-before-commit. Separately, the Neon CLI is now installable as neon from npm (the old neonctl name still works), and adds neon config init for declarative branch configs plus neon status for offline branch checks.
Supabase released CLI v2.109.0 with a critical fix: migration commands now correctly apply pipeline-incompatible SQL statements like CREATE INDEX CONCURRENTLY and VACUUM. The release also adds db reset --sql-paths for CI workflows and ports migration commands to native TypeScript. The Supabase Client SDK v2.109.0 introduced a Realtime postgres_changes filter builder with additional operators, plus purgeCache for Storage buckets and objects. SDK v2.110.0 drops Node.js 20 support, requiring Node.js 22 or later.
Auth and schema refinements
Supabase Auth v2.192.0 ships a per-provider custom_claims_allowlist for custom OAuth providers, giving finer control over which claims appear in tokens. Prisma Next published setup guides for new and existing PostgreSQL and MongoDB projects, and extended the schema language with scalar list fields and MongoDB enum support. Supabase's OpenCode integration launched, letting AI coding agents authenticate with /supabase to access account and project management APIs.
Releases covered
- PlanetScale Query Insights adds tag-based filtering for Vitess databases
- PlanetScale Query Insights now shows vindex usage for Vitess databases
- PlanetScale adds VTTablet and MySQL configuration to dashboard
- PlanetScale MCP server adds replica targeting and Postgres RLS detection
- PlanetScale Terraform provider v1.3.0 adds password-free Postgres role management
- Supabase CLI v2.109.0 fixes pipeline-incompatible migration statements
- Supabase Client SDK v2.109.0 adds Realtime filter builder and Storage cache purge
- Supabase Client SDK v2.110.0 drops Node.js 20 support
- Supabase Auth v2.192.0 adds per-provider custom claims allowlist for OAuth
- Supabase OpenCode integration launches with account and project management APIs