feat: include filters in dashboard exports

.changeset/add-accesskey-index.md

@@ -0,0 +1,5 @@
+---
+"@hyperdx/api": minor
+---
+
+Add unique MongoDB index on accessKey field in User model to eliminate full collection scans during API key authentication. This could cause startup failures if any existing users share duplicate accessKey values.

.changeset/good-swans-hammer.md

@@ -0,0 +1,5 @@
+---
+"@hyperdx/common-utils": patch
+---
+
+Adding filters to dashboard exports. Implemented validation on dashboard imports to catch potential issues with generated JSON or manually tweaked JSON.

.changeset/mcp-dashboard-improvements.md

@@ -0,0 +1,21 @@
+---
+'@hyperdx/api': patch
+---
+
+feat(mcp): add patch_dashboard, get_dashboard_tile, search_dashboards tools
+
+Add three new MCP dashboard tools for granular operations:
+
+- `hyperdx_get_dashboard_tile` — retrieve a single tile by tileId
+- `hyperdx_patch_dashboard` — update name/tags and/or replace one tile
+  without resubmitting the full dashboard
+- `hyperdx_search_dashboards` — search by name and/or tags
+
+Fix empty parameter schema on patch/search tools caused by Zod
+`.refine()` wrapping. Document Lucene substring matching limitations
+prominently in tool descriptions and query guide prompt.
+
+**Breaking (minor):** Tile `name` on `hyperdx_save_dashboard` now requires
+at least 1 character (`.min(1)`). Previously empty string `""` was accepted
+and silently persisted as a blank title. Callers sending `name: ""` will
+now receive a validation error.

.changeset/page-header-z-index-below-drawers.md

@@ -0,0 +1,20 @@
+---
+'@hyperdx/app': patch
+---
+
+fix(z-index): keep sticky header below drawers and drawers above the fullscreen tile modal
+
+Two related z-index regressions:
+
+- `PageHeader` was pinned at `z-index: 100`, but app drawers opt into a
+  much lower stack via `ZIndexContext` (`contextZIndex + 10`, so a
+  top-level drawer renders at `z-index: 10`). The sticky header therefore
+  floated above the drawer overlay. The header now sits at `z-index: 2` so
+  drawer overlays reliably cover the page chrome while the header still
+  wins against normal scrolling content.
+- `FullscreenPanelModal` used Mantine's default modal z-index (`200`) and
+  didn't propagate it through `ZIndexContext`. Clicking a row in a
+  fullscreen search tile opened a `DBRowSidePanel` drawer at `z-index: 10`
+  that was hidden behind the modal. The modal now follows the existing
+  `contextZIndex + 10` pattern and wraps its children in a
+  `ZIndexContext.Provider`, so child drawers stack on top of it.

.changeset/page-migrate-dashboard.md

@@ -0,0 +1,7 @@
+---
+'@hyperdx/app': patch
+---
+
+chore: migrate the custom Dashboard page to shared `PageLayout` / `PageHeader`. Breadcrumbs, the editable dashboard name, dashboard actions (Favorite, Tags, Menu), and the "Created by … Updated …" meta now live in a single page header, while the query toolbar (SQL/Lucene WHERE, time range, granularity, Live, refresh, edit filters, Run) is pinned to the top of the scroll container as a dedicated sticky row — the chrome above scrolls away and only the toolbar follows the user. The "Updated …" meta moves to the right side of the breadcrumbs row instead of sitting as a separate body line.
+
+`PageHeader` gains a `stickyRow` slot that any page can use to declare a single row that should be the only pinned element, with the rest of the header treated as scrolling chrome. Other pages are unaffected — a `PageHeader` without `stickyRow` keeps the existing fully-sticky behavior.

.changeset/rename-mcp-tools-to-clickstack.md

@@ -0,0 +1,10 @@
+---
+"@hyperdx/api": patch
+"@hyperdx/common-utils": patch
+---
+
+refactor(mcp): rename all MCP tool prefixes from `hyperdx_` to `clickstack_`
+
+Rename the MCP server name from `hyperdx` to `clickstack` and update all 19
+tool names (e.g. `hyperdx_search` → `clickstack_search`), along with
+descriptions, prompts, error messages, and test references.

.changeset/source-chip-kebab-menu.md

@@ -0,0 +1,5 @@
+---
+'@hyperdx/app': patch
+---
+
+feat(source-picker): chip + kebab menu UX

.github/actions/e2e-setup/action.yml

@@ -0,0 +1,91 @@
+name: E2E Setup
+description: |
+  Shared setup for E2E test workflows: sets the HDX_E2E_* port env vars and
+  common Playwright env vars, then runs Node.js setup, dependency install,
+  build, Playwright browser install, and starts the E2E Docker Compose stack
+  (MongoDB + ClickHouse) with health checks. The caller is responsible for
+  `actions/checkout`, running Playwright, uploading artifacts, and tearing
+  down the Docker stack.
+runs:
+  using: composite
+  steps:
+    # E2E port configuration (slot 0 defaults — must match scripts/test-e2e.sh)
+    - name: Set E2E environment variables
+      shell: bash
+      run: |
+        MONGO_PORT=21100
+        echo "HDX_E2E_SLOT=0" >> $GITHUB_ENV
+        echo "HDX_E2E_OPAMP_PORT=20320" >> $GITHUB_ENV
+        echo "HDX_E2E_CH_PORT=20500" >> $GITHUB_ENV
+        echo "HDX_E2E_CH_NATIVE_PORT=20600" >> $GITHUB_ENV
+        echo "HDX_E2E_API_PORT=21000" >> $GITHUB_ENV
+        echo "HDX_E2E_MONGO_PORT=$MONGO_PORT" >> $GITHUB_ENV
+        echo "HDX_E2E_APP_LOCAL_PORT=21200" >> $GITHUB_ENV
+        echo "HDX_E2E_APP_PORT=21300" >> $GITHUB_ENV
+        echo "MONGO_URI=mongodb://localhost:$MONGO_PORT/hyperdx-e2e" >> $GITHUB_ENV
+        echo "E2E_FULLSTACK=true" >> $GITHUB_ENV
+        echo "E2E_UNIQUE_USER=true" >> $GITHUB_ENV
+        echo "E2E_API_HEALTH_CHECK_MAX_RETRIES=60" >> $GITHUB_ENV
+
+    - name: Setup Node.js
+      uses: actions/setup-node@v6
+      with:
+        node-version-file: .nvmrc
+        cache-dependency-path: yarn.lock
+        cache: yarn
+
+    - name: Install dependencies
+      shell: bash
+      run: yarn install
+
+    - name: Build dependencies
+      shell: bash
+      run: npx nx run-many -t ci:build
+
+    - name: Install Playwright browsers
+      shell: bash
+      run: cd packages/app && npx playwright install --with-deps chromium
+
+    - name: Start E2E Docker Compose
+      shell: bash
+      run: |
+        # Pre-pull images with retries to handle transient Docker Hub timeouts
+        for attempt in 1 2 3; do
+          if docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml pull; then
+            echo "Docker images pulled successfully on attempt $attempt"
+            break
+          fi
+          if [ "$attempt" -eq 3 ]; then
+            echo "Failed to pull Docker images after 3 attempts"
+            exit 1
+          fi
+          echo "Docker pull failed (attempt $attempt/3), retrying in 10s..."
+          sleep 10
+        done
+        docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml up -d
+        echo "Waiting for MongoDB..."
+        for i in $(seq 1 30); do
+          if docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml exec -T db mongosh --quiet --eval "db.adminCommand({ping:1})" >/dev/null 2>&1; then
+            echo "MongoDB is ready"
+            break
+          fi
+          if [ "$i" -eq 30 ]; then
+            echo "MongoDB failed to become ready after 30 seconds"
+            exit 1
+          fi
+          echo "Waiting for MongoDB... ($i/30)"
+          sleep 1
+        done
+        echo "Waiting for ClickHouse..."
+        for i in $(seq 1 60); do
+          if curl -sf http://localhost:${HDX_E2E_CH_PORT}/ping >/dev/null 2>&1; then
+            echo "ClickHouse is ready"
+            break
+          fi
+          if [ "$i" -eq 60 ]; then
+            echo "ClickHouse failed to become ready after 60 seconds"
+            exit 1
+          fi
+          echo "Waiting for ClickHouse... ($i/60)"
+          sleep 1
+        done

.github/actions/e2e-teardown/action.yml

@@ -0,0 +1,35 @@
+name: E2E Teardown
+description: |
+  Upload Playwright artifacts and stop the E2E Docker Compose stack.
+  Always runs (even on failure) so containers are cleaned up and reports
+  are available for failed runs.
+inputs:
+  artifact-suffix:
+    description: >
+      Suffix appended to artifact names, e.g. "1" → playwright-report-1.
+    required: true
+runs:
+  using: composite
+  steps:
+    - name: Upload Playwright report
+      uses: actions/upload-artifact@v7
+      if: always()
+      with:
+        name: playwright-report-${{ inputs.artifact-suffix }}
+        path: packages/app/playwright-report/
+        retention-days: 30
+
+    - name: Upload test results
+      uses: actions/upload-artifact@v7
+      if: always()
+      with:
+        name: test-results-${{ inputs.artifact-suffix }}
+        path: packages/app/test-results/
+        retention-days: 30
+
+    - name: Stop E2E containers
+      if: always()
+      shell: bash
+      run:
+        docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml
+        down -v

.github/workflows/e2e-tests.yml

@@ -14,109 +14,20 @@ jobs:
       matrix:
         shard: [1, 2, 3, 4]
 
-    # E2E port configuration (slot 0 defaults — must match scripts/test-e2e.sh)
-    env:
-      HDX_E2E_SLOT: '0'
-      HDX_E2E_OPAMP_PORT: '20320'
-      HDX_E2E_CH_PORT: '20500'
-      HDX_E2E_CH_NATIVE_PORT: '20600'
-      HDX_E2E_API_PORT: '21000'
-      HDX_E2E_MONGO_PORT: '21100'
-      HDX_E2E_APP_LOCAL_PORT: '21200'
-      HDX_E2E_APP_PORT: '21300'
-
     steps:
       - name: Checkout
         uses: actions/checkout@v6
 
-      - name: Setup Node.js
-        uses: actions/setup-node@v6
-        with:
-          node-version-file: '.nvmrc'
-          cache-dependency-path: 'yarn.lock'
-          cache: 'yarn'
-
-      - name: Install dependencies
-        run: yarn install
-
-      - name: Build dependencies
-        run: npx nx run-many -t ci:build
-
-      - name: Install Playwright browsers
-        run: cd packages/app && npx playwright install --with-deps chromium
-
-      - name: Start E2E Docker Compose
-        run: |
-          # Pre-pull images with retries to handle transient Docker Hub timeouts
-          for attempt in 1 2 3; do
-            if docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml pull; then
-              echo "Docker images pulled successfully on attempt $attempt"
-              break
-            fi
-            if [ "$attempt" -eq 3 ]; then
-              echo "Failed to pull Docker images after 3 attempts"
-              exit 1
-            fi
-            echo "Docker pull failed (attempt $attempt/3), retrying in 10s..."
-            sleep 10
-          done
-          docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml up -d
-          echo "Waiting for MongoDB..."
-          for i in $(seq 1 30); do
-            if docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml exec -T db mongosh --quiet --eval "db.adminCommand({ping:1})" >/dev/null 2>&1; then
-              echo "MongoDB is ready"
-              break
-            fi
-            if [ "$i" -eq 30 ]; then
-              echo "MongoDB failed to become ready after 30 seconds"
-              exit 1
-            fi
-            echo "Waiting for MongoDB... ($i/30)"
-            sleep 1
-          done
-          echo "Waiting for ClickHouse..."
-          for i in $(seq 1 60); do
-            if curl -sf http://localhost:${HDX_E2E_CH_PORT}/ping >/dev/null 2>&1; then
-              echo "ClickHouse is ready"
-              break
-            fi
-            if [ "$i" -eq 60 ]; then
-              echo "ClickHouse failed to become ready after 60 seconds"
-              exit 1
-            fi
-            echo "Waiting for ClickHouse... ($i/60)"
-            sleep 1
-          done
+      - name: E2E setup
+        uses: ./.github/actions/e2e-setup
 
       - name: Run Playwright tests (full-stack mode)
-        env:
-          E2E_FULLSTACK: 'true'
-          E2E_UNIQUE_USER: 'true'
-          E2E_API_HEALTH_CHECK_MAX_RETRIES: '60'
-          MONGO_URI:
-            mongodb://localhost:${{ env.HDX_E2E_MONGO_PORT }}/hyperdx-e2e
         run: |
           cd packages/app
           yarn test:e2e --shard=${{ matrix.shard }}/4
 
-      - name: Upload Playwright report
-        uses: actions/upload-artifact@v7
-        if: always()
-        with:
-          name: playwright-report-${{ matrix.shard }}
-          path: packages/app/playwright-report/
-          retention-days: 30
-
-      - name: Upload test results
-        uses: actions/upload-artifact@v7
+      - name: E2E teardown
         if: always()
+        uses: ./.github/actions/e2e-teardown
         with:
-          name: test-results-${{ matrix.shard }}
-          path: packages/app/test-results/
-          retention-days: 30
-
-      - name: Stop E2E containers
-        if: always()
-        run:
-          docker compose -p e2e-0 -f packages/app/tests/e2e/docker-compose.yml
-          down -v
+          artifact-suffix: ${{ matrix.shard }}

AGENTS.md

@@ -71,19 +71,6 @@ and lint issues across all packages. Pre-commit hooks handle this when
 committing, but if you finish edits without committing, run `yarn lint:fix`
 before stopping.
 
-## Before Writing New Utility Functions
-
-Before implementing a new parsing, formatting, or transformation function,
-search for existing functions that already do the same thing or something close.
-Common places to check:
-
-- `packages/common-utils/src/core/metadata.ts` — key/column parsing utilities
-  (e.g. `parseKeyPath`)
-- `packages/common-utils/src/queryParser.ts` — Lucene parsing (`parse`,
-  `decodeSpecialTokens`), AST type guards, serialization helpers
-- `packages/common-utils/src/filters.ts` — filter state serialization and
-  parsing (`filtersToQuery`, `parseLuceneFilter`)
-
 ## Key Principles
 
 1. **Multi-tenancy**: All data is scoped to `Team` - ensure proper filtering