Zero 1.9

Query Correctness and Reliability

Installation

Copynpm install @rocicorp/zero@1.9

Overview

Zero 1.9 improves query/mutation correctness and reliability.

Features

• Litestream v5 restores: The official image now uses Litestream 0.5.15 for restores by default. It can restore legacy WAL or LTX backups, while Zero 1.9 continues writing legacy backups. Legacy snapshots now retain the previous generation for six additional hours, preventing cleanup during an active restore. (#6260, #6267)
• End-to-end serving lag: New metrics measure completed replicated work from the upstream transaction commit through zero-cache sync output. Upstream clock-skew estimate identifies measurements biased by clock differences. (#6312)

Performance

First Mutation Latency

On the first mutation, Zero Server fetches and caches PostgreSQL schema metadata. In benchmarks, that full request is 2.7x faster (done in #6292, thanks @diegopereira99!). This is most noticeable with cold-starts in serverless environments like AWS Lambda.

Fixes

• Ordered queries now paginate and maintain windows correctly when cursor fields contain NULL, including compound tie-break fields and reverse walks. (thanks @YevheniiKotyrlo!)
• Rebuilt queries now deliver changed rows instead of occasionally leaving clients with stale results.
• Queries no longer drop rows or emit invalid SQL when given an inapplicable scalar hint, and scalar NOT EXISTS now handles empty or NULL results.
• Schema construction, CRUD mutators, and materialized views now preserve a key named __proto__ as user data. (thanks @tjenkinson!)
• SQLite statement caches now retain at most 1,000 idle entries each, terminated client groups release custom-query timers and caches, and large replica transactions can spill dirty pages to WAL instead of retaining the complete write set in native memory.
• Missing replication-lag reports are retried and total_lag no longer grows when reports stop arriving, while serving-lag metrics exclude disconnected or not-yet-validated client groups.
• zero-cache now recovers from half-open PostgreSQL sockets, including over TLS, and the official image applies the bundled postgres.js disconnect patch.
• With PostgreSQL wal_sender_timeout=0, replication no longer enters a continuous reconnect loop. See WAL Sender Timeout.
• Client connection attempts now time out across setup and the server handshake, abandon late sockets, and retry normally. Reconnect confirmations no longer produce false slow-query warnings or inflated materialization metrics.
• Different integration versions in a pnpm workspace no longer create peer-qualified duplicate copies of @rocicorp/zero, fixing cross-package type and module-augmentation failures.
• Replicated PostgreSQL type and nullability changes now preserve compound-index column order in SQLite replicas. To repair an affected replica, resync it from Postgres or recreate the PostgreSQL index.
• Expected schema and replica resets now log warnings instead of errors, and zero-cache skips Litestream restore when backups are not configured. (thanks @asterikx!)
• Server CRUD updates and upserts no longer assign primary-key columns, avoiding PostgreSQL locks that could block concurrent foreign-key inserts. (thanks @shayonj!)
• Mutation and query API calls now retry all 5xx responses using the existing four-attempt limit and backoff; 4xx responses still fail without retry. (thanks @shayonj!)
• SQLite corruption failures now log bounded replica and integrity diagnostics and flush logs before exit. Oversized replication updates now identify the transaction, affected column, and value type without logging the value.

Breaking Changes

Existing Primary-Key Inserts

insert now succeeds without changing the row when its Zero primary key already exists; before 1.9, the server returned an error. A successful insert does not necessarily prove that a row was created.