PlanetScale

Recently shipped backup webhook events, Postgres restore-point branches, and a large expansion of CLI and MCP server operations.

Backups gained webhook events and deletion protection. backup.succeeded and backup.failed webhook events now fire for both Vitess and Postgres databases. The CLI's pscale backup update can mark a backup protected so it can't be deleted until protection is turned off.

Postgres branches can be created from a point in time. pscale branch create accepts --restore-point with --restore or --from to build a branch from a specific backup or the backup covering a timestamp. Postgres only; can't be combined with --seed-data.

The MCP server can now answer questions about query tags, errors, and Postgres logs. New tools read query tags from Insights, pull failing queries grouped by error fingerprint, and fetch Postgres server logs filtered by level, time range, role, or raw LogsQL. Ask which route tag spends the most database time, or what failed in the last hour, from your editor.

CLI coverage expanded across billing, orgs, and operations. New commands list invoices and line items, update payment methods via Stripe Checkout, update org billing and spend-alert settings, manage teams and members, show service tokens, list database regions, and run branch maintenance. Insights gained per-query detail, schema recommendation DDL, and query-pattern report management; metrics gained dedicated queries, tables, tablets, and tags commands; Postgres switchovers can be listed and shown by ID.

The dashboard can now manage live connections. The Connections view shows active Postgres and Vitess sessions with search, sort, and filters, and lets you cancel a running query or terminate its connection. A new Prefer instant admin setting defaults Vitess deploy requests to instant DDL when available — CLI and API deploys still require the explicit flag.