2.3.0 rc

.agent/notes/driver-test-progress.md

@@ -68,3 +68,11 @@ Scope: DB driver tests only
 - 2026-05-08 16:52 PDT raw-websocket [native/local/bare]: PASS (16 passed, 14.0s). Covers raw websocket callback tracking.
 - 2026-05-08 16:52 PDT actor-conn-state [native/local/bare]: PASS (11 passed, 10.9s). Covers connection lifecycle and disconnect accounting through the work registry.
 - 2026-05-08 16:52 PDT actor-sleep-db [native/local/bare]: PASS (26 passed, 70.9s). Covers DB close timing during sleep shutdown and waitUntil state persistence.
+- 2026-05-09 04:48 PDT DB TESTS RERUN STARTED [native only] - validating async websocket close handler behavior.
+- 2026-05-09 04:48 PDT actor-db rerun [native]: PASS (13 passed, 104 skipped, 18.3s).
+- 2026-05-09 04:48 PDT actor-db-raw rerun [native]: PASS (5 passed, 40 skipped, 4.9s).
+- 2026-05-09 04:48 PDT actor-db-pragma-migration rerun [native]: PASS (4 passed, 32 skipped, 4.3s).
+- 2026-05-09 04:48 PDT actor-sleep-db rerun [native]: PASS (26 passed, 208 skipped, 63.8s). Includes `async websocket close handler can use c.db before sleep completes` and `async websocket addEventListener close handler can use c.db before sleep completes`.
+- 2026-05-09 04:48 PDT actor-db-stress rerun [native]: PASS (5 passed, 40 skipped, 31.3s).
+- 2026-05-09 04:48 PDT actor-db-init-order rerun [native]: PASS (6 passed, 48 skipped, 6.3s).
+- 2026-05-09 04:48 PDT DB TESTS RERUN COMPLETE [native only] - 6/6 DB file groups passed. Async close handler tests included.

.claude/reference/content-frontmatter.md

@@ -23,3 +23,5 @@ Required fields:
 Optional fields:
 
 - `keywords` (string array)
+- `image` (`true` or `{ format: string }`) — hero image flag. Presence means the post has a hero image in R2 at `website/blog/{post-slug}/image.{format}`; use `image: true` for the default `image.png` or `image: { format: "gif" }` for another extension. The URL is derived from the slug and dimensions are fixed (2:1), so do not write `src`, `width`, or `height`. Resolved by `website/src/lib/postImage.ts`.
+- `unpublished` (boolean)

.gitattributes

@@ -1,12 +1,15 @@
+# Do not add Git LFS filters here. jj can snapshot raw working-tree bytes
+# without running Git LFS clean filters, so large public assets belong in R2.
+
 * text eol=lf
 *.png binary
 *.jpg binary
 *.jpeg binary
-*.gif filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tar.gz filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-engine/packages/test-snapshot-gen/snapshots/** filter=lfs diff=lfs merge=lfs -text
+*.gif binary
+*.tar binary
+*.tar.gz binary
+*.tgz binary
+*.sst binary
 
 **/Cargo.lock linguist-generated=true
 

.github/workflows/publish.yaml

@@ -217,42 +217,42 @@ jobs:
           if-no-files-found: error
 
   # ---------------------------------------------------------------------------
-  # build-wasm — wasm package artifact built in parallel with native artifacts
+  # build-wasm — disabled
   # ---------------------------------------------------------------------------
-  build-wasm:
-    needs: [context]
-    name: "Build rivetkit-wasm"
-    runs-on: depot-ubuntu-24.04-8
-    permissions:
-      contents: read
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          lfs: ${{ needs.context.outputs.trigger == 'release' }}
-      - run: corepack enable
-      - uses: actions/setup-node@v4
-        with:
-          node-version: "22"
-          cache: pnpm
-      - uses: actions-rust-lang/setup-rust-toolchain@v1
-        with:
-          toolchain: stable
-          target: wasm32-unknown-unknown
-          rustflags: ""
-      - uses: Swatinem/rust-cache@v2
-        with:
-          shared-key: "rivetkit-wasm-publish"
-          cache-on-failure: true
-      - name: Install wasm package dependencies
-        run: pnpm install --frozen-lockfile --filter=@rivetkit/rivetkit-wasm
-      - name: Build wasm package
-        run: pnpm --filter=@rivetkit/rivetkit-wasm build
-      - name: Upload wasm package artifact
-        uses: actions/upload-artifact@v4
-        with:
-          name: wasm-package
-          path: rivetkit-typescript/packages/rivetkit-wasm/pkg
-          if-no-files-found: error
+  # build-wasm:
+  #   needs: [context]
+  #   name: "Build rivetkit-wasm"
+  #   runs-on: depot-ubuntu-24.04-8
+  #   permissions:
+  #     contents: read
+  #   steps:
+  #     - uses: actions/checkout@v4
+  #       with:
+  #         lfs: ${{ needs.context.outputs.trigger == 'release' }}
+  #     - run: corepack enable
+  #     - uses: actions/setup-node@v4
+  #       with:
+  #         node-version: "22"
+  #         cache: pnpm
+  #     - uses: actions-rust-lang/setup-rust-toolchain@v1
+  #       with:
+  #         toolchain: stable
+  #         target: wasm32-unknown-unknown
+  #         rustflags: ""
+  #     - uses: Swatinem/rust-cache@v2
+  #       with:
+  #         shared-key: "rivetkit-wasm-publish"
+  #         cache-on-failure: true
+  #     - name: Install wasm package dependencies
+  #       run: pnpm install --frozen-lockfile --filter=@rivetkit/rivetkit-wasm
+  #     - name: Build wasm package
+  #       run: pnpm --filter=@rivetkit/rivetkit-wasm build
+  #     - name: Upload wasm package artifact
+  #       uses: actions/upload-artifact@v4
+  #       with:
+  #         name: wasm-package
+  #         path: rivetkit-typescript/packages/rivetkit-wasm/pkg
+  #         if-no-files-found: error
 
   # ---------------------------------------------------------------------------
   # docker-images — per-arch runtime images pushed to Docker Hub
@@ -319,12 +319,11 @@ jobs:
   # publish — npm publish + R2 upload + Docker manifest + release tail
   # ---------------------------------------------------------------------------
   publish:
-    needs: [context, build, build-wasm, docker-images]
+    needs: [context, build, docker-images]
     name: "Publish"
     if: |
       !cancelled() &&
       needs.build.result == 'success' &&
-      needs.build-wasm.result == 'success' &&
       needs.docker-images.result == 'success'
     runs-on: depot-ubuntu-24.04-8
     permissions:
@@ -364,17 +363,6 @@ jobs:
           path: engine-artifacts
           pattern: engine-*
           merge-multiple: true
-      - name: Download wasm package artifact
-        uses: actions/download-artifact@v4
-        with:
-          name: wasm-package
-          path: rivetkit-typescript/packages/rivetkit-wasm/pkg
-      - name: Validate wasm package artifact
-        run: |
-          test -f rivetkit-typescript/packages/rivetkit-wasm/pkg/rivetkit_wasm.js
-          test -f rivetkit-typescript/packages/rivetkit-wasm/pkg/rivetkit_wasm.d.ts
-          test -f rivetkit-typescript/packages/rivetkit-wasm/pkg/rivetkit_wasm_bg.wasm
-
       - name: Place native binaries in platform packages
         run: |
           NATIVE_DIR=rivetkit-typescript/packages/rivetkit-napi

CLAUDE.md

@@ -24,6 +24,7 @@ Design constraints, invariants, and reference commands for the Rivet monorepo. F
 
 - Avoid raw `f64` fields in vbare protocol schemas that use hashable maps; generated Rust derives `Eq`/`Hash`, so encode floats as fixed bytes or an ordered wrapper.
 - Version converters must manually map fields between versions; never use serialize-deserialize round trips such as `transcode_version` or `serde_bare::to_vec` plus `from_slice`.
+- RivetKit client/server protocol compatibility assumes the server/runtime is newer than the client; clients send their latest request protocol version, and servers handle older-client compatibility and negotiation.
 
 When talking about "Rivet Actors" make sure to capitalize "Rivet Actor" as a proper noun and lowercase "actor" as a generic noun.
 
@@ -79,6 +80,14 @@ docker-compose up -d
 
 **Never push to `main` unless explicitly specified by the user.**
 
+## Assets
+
+- Large public dashboard and website media belongs in the `rivet-assets` R2 bucket, not Git.
+- Use object keys shaped like `dashboard/{group}/{asset-name}` for dashboard media and `website/blog/{post}/{asset-name}` for blog or changelog post media.
+- Upload with `op://Engineering/rivet-assets R2 Upload/{username,password}` and `aws s3 cp <file> s3://rivet-assets/<key> --endpoint-url https://2a94c6a0ced8d35ea63cddc86c2681e7.r2.cloudflarestorage.com`.
+- For blog or changelog hero media, upload the file to R2 as `website/blog/{post-slug}/image.{ext}` and set frontmatter `image: true` (or `image: { format: "gif" }` for a non-png). The URL is derived from the slug and dimensions are fixed at a 2:1 ratio, so do not write the absolute URL, width, or height. Resolved by `website/src/lib/postImage.ts`.
+- Do not use Git LFS in this repo; jj can snapshot raw working-tree bytes.
+
 ## Frontend Visual Changes
 
 - For any frontend visual change, use the `agent-browser` skill to view the result in a browser instead of working blind. If it is not installed, prompt the user to install it.

README.md

@@ -142,7 +142,7 @@ One primitive that adapts to agents, workflows, collaboration, and more.
 Powerful debugging and monitoring tools from local development to production at scale.
 
 <p align="center">
-  <img src="./.github/media/screenshots/inspector.png" alt="Rivet Inspector" width="800" />
+  <img src="https://assets.rivet.dev/github/media/screenshots/inspector.png" alt="Rivet Inspector" width="800" />
 </p>
 
 - **SQLite Viewer** — Browse and query your actor's SQLite database in real-time

docs-internal/engine/SQLITE_METRICS.md

@@ -6,8 +6,8 @@
 - `sqlite_commit_dirty_page_count{path}`: Histogram of dirty page counts per commit path.
 - `sqlite_commit_dirty_bytes{path}`: Histogram of raw dirty-page bytes per commit path.
 - `sqlite_udb_ops_per_commit{path}`: Histogram of UniversalDB operations per commit path.
-- `sqlite_commit_envoy_dispatch_duration_seconds`: Pegboard-envoy histogram for websocket frame arrival to `depot` dispatch.
-- `sqlite_commit_envoy_response_duration_seconds`: Pegboard-envoy histogram for `depot` return to websocket response send.
+- `envoy_sqlite_commit_dispatch_duration_seconds`: Pegboard-envoy histogram for websocket frame arrival to `depot` dispatch.
+- `envoy_sqlite_commit_response_duration_seconds`: Pegboard-envoy histogram for `depot` return to websocket response send.
 - `sqlite_commit_phases`: Actor inspector labeled timing metric exposed from `/inspector/metrics`. Values are `request_build`, `serialize`, `transport`, and `state_update`.
 
 ## Scrape Points
@@ -23,8 +23,8 @@
 
 ## Diagnosis
 
-- High `decode_request` or `sqlite_commit_envoy_dispatch_duration_seconds` usually means envoy-side validation or actor lookup is slow before storage work starts.
+- High `decode_request` or `envoy_sqlite_commit_dispatch_duration_seconds` usually means envoy-side validation or actor lookup is slow before storage work starts.
 - High `meta_read` or `pidx_read` points at UniversalDB read pressure or cache misses.
 - High `ltx_encode` means commit encoding and compression are doing real work. Check dirty page counts and raw dirty bytes together.
-- High `udb_write`, `meta_write`, or `sqlite_commit_envoy_response_duration_seconds` points at write-path latency after encode.
+- High `udb_write`, `meta_write`, or `envoy_sqlite_commit_response_duration_seconds` points at write-path latency after encode.
 - A healthy actor should show non-zero `sqlite_commit_phases` totals after commits in `/inspector/metrics`. If SQL runs but those timings stay zero, the native VFS metrics path is broken.

docs-internal/engine/TEST_SNAPSHOTS.md

@@ -9,7 +9,7 @@ The `test-snapshot-gen` crate (`engine/packages/test-snapshot-gen/`) provides:
 1. **A binary** (`test-snapshot-gen`) that runs scenarios to generate RocksDB snapshots.
 2. **A library** (`test_snapshot`) that loads those snapshots into test infrastructure.
 
-Snapshots capture the entire UDB state (epoxy, gasoline, pegboard, etc.) for each replica in a multi-node cluster. They are stored as raw RocksDB checkpoint directories tracked by git LFS.
+Snapshots capture the entire UDB state (epoxy, gasoline, pegboard, etc.) for each replica in a multi-node cluster. They are stored as raw RocksDB checkpoint directories checked in as normal fixture files.
 
 ## Generating Snapshots
 
@@ -140,6 +140,6 @@ let replica_paths = load_snapshot("epoxy-v1", test_id).unwrap();
 4. A `metadata.json` file is written with the commit hash, branch name, and timestamp.
 5. The test loader copies the checkpoint directory to `$TMPDIR/rivet-test-{test_id}-{dc_label}`, which is the same path that `rivet_test_deps::setup_single_datacenter` creates. Since the directory already exists with data, the RocksDB driver opens it and finds the pre-populated state.
 
-## Git LFS
+## Git
 
-All files under `engine/packages/test-snapshot-gen/snapshots/` are tracked by git LFS (configured in `.gitattributes`). Make sure git LFS is installed before committing snapshots.
+All files under `engine/packages/test-snapshot-gen/snapshots/` are checked in directly. Keep scenarios small enough that generated snapshots stay lightweight.

engine/CLAUDE.md

@@ -43,7 +43,7 @@ When changing a versioned VBARE schema, follow the existing migration pattern.
 
 ## Test snapshots
 
-Use `test-snapshot-gen` to generate and load RocksDB snapshots of the full UDB KV store for migration and integration tests. Scenarios produce per-replica RocksDB checkpoints stored under `engine/packages/test-snapshot-gen/snapshots/` (git LFS tracked). In tests, use `test_snapshot::SnapshotTestCtx::from_snapshot("scenario-name")` to boot a cluster from snapshot data. See `docs-internal/engine/TEST_SNAPSHOTS.md` for the full guide.
+Use `test-snapshot-gen` to generate and load RocksDB snapshots of the full UDB KV store for migration and integration tests. Scenarios produce per-replica RocksDB checkpoints stored under `engine/packages/test-snapshot-gen/snapshots/` as normal checked-in fixture files. In tests, use `test_snapshot::SnapshotTestCtx::from_snapshot("scenario-name")` to boot a cluster from snapshot data. See `docs-internal/engine/TEST_SNAPSHOTS.md` for the full guide.
 
 ## Engine test flakes
 

engine/packages/config/src/config/pegboard.rs

@@ -159,7 +159,7 @@ pub struct Pegboard {
 
 impl Pegboard {
 	pub fn base_retry_timeout(&self) -> usize {
-		self.base_retry_timeout.unwrap_or(2000)
+		self.base_retry_timeout.unwrap_or(2_000)
 	}
 
 	pub fn actor_allocation_threshold(&self) -> i64 {
@@ -177,7 +177,7 @@ impl Pegboard {
 	}
 
 	pub fn actor_retry_duration_threshold(&self) -> i64 {
-		self.actor_retry_duration_threshold.unwrap_or(300_000)
+		self.actor_retry_duration_threshold.unwrap_or(5 * 60 * 1000)
 	}
 
 	pub fn retry_reset_duration(&self) -> i64 {
@@ -202,7 +202,7 @@ impl Pegboard {
 	}
 
 	pub fn serverless_base_retry_timeout(&self) -> usize {
-		self.serverless_base_retry_timeout.unwrap_or(2000)
+		self.serverless_base_retry_timeout.unwrap_or(2_000)
 	}
 
 	pub fn serverless_retry_reset_duration(&self) -> i64 {
@@ -237,7 +237,7 @@ impl Pegboard {
 
 	pub fn gateway_response_start_timeout_ms(&self) -> u64 {
 		self.gateway_response_start_timeout_ms
-			.unwrap_or(5 * 60 * 1000) // 5 minutes
+			.unwrap_or(5 * 60 * 1000)
 	}
 
 	pub fn gateway_update_ping_interval_ms(&self) -> u64 {