Replace per-cell benchmarks with shared native database sessions

Replace the 1,898-worker comparison pipeline with two measurement jobs per
database, using one live Aspire topology for separate native TUnit cases.
Measure the initial million-record load and bounded mixed operation streams
without warmup, reseeding or per-case infrastructure restarts.

Remove superseded planners, dispatchers, runners, report schemas/readers and
their exclusive fixtures. Preserve native adapters, correctness oracles,
historical evidence and required functional RF3 qualification.

Build and authenticate a common compiled runner bundle before fan-out. Retain
original per-case outcomes, native reports, source/image/runtime identities,
failed/null results and explicit native capability/topology unavailability.
Publication requires a successful, complete authenticated cohort.

This is a source checkpoint. The new comparison and site tests, real Linux
sessions and complete model-family qualification remain open. Initial session
coverage contains seven document cases; it does not qualify the vector, graph,
queue, event or time-series families.

Author: KSemenenko

.github/workflows/Features/BenchmarkComparisons/FunctionalHeavyLoadSetup/AGENTS.md

@@ -0,0 +1,3 @@
+# FunctionalHeavyLoadSetup
+
+Prepare the existing KeyLoad functional RF3 million-record ingestion gate with its exact-source images and native runner. Follow root and workflow policies. This action runs no workload and produces no measurements. Preserve the original native admission/workload steps and original result upload as top-level Benchmarks job steps for authenticated provider accounting. No database reset or additional measurement matrix.

.github/workflows/Features/BenchmarkComparisons/FunctionalHeavyLoadSetup/action.yml

@@ -0,0 +1,38 @@
+name: Functional RF3 heavy-load setup
+description: Prepare source-bound images and the existing native functional RF3 runner.
+inputs:
+  bundle-artifact-id:
+    description: Same-source immutable Docker image bundle
+    required: true
+runs:
+  using: composite
+  steps:
+    - name: Set up .NET
+      uses: actions/setup-dotnet@a98b56852c35b8e3190ac28c8c2271da59106c68
+      with:
+        global-json-file: global.json
+    - name: Check Docker
+      shell: bash
+      run: docker version
+    - name: Download exact source Docker images
+      uses: actions/download-artifact@37930b1c2abaa49bbe596cd826c3c89aef350131
+      with:
+        artifact-ids: ${{ inputs.bundle-artifact-id }}
+        path: ${{ runner.temp }}/keyload-image-bundle
+    - name: Load and verify Docker images
+      id: images
+      shell: bash
+      run: node scripts/Features/BenchmarkComparisons/import-images.mjs --bundle="$RUNNER_TEMP/keyload-image-bundle"
+    - name: Configure exact source KeyLoad image
+      env:
+        KEYLOAD_SERVER_IMAGE: ${{ steps.images.outputs.server-image }}
+      shell: bash
+      run: |
+        test -n "$KEYLOAD_SERVER_IMAGE"
+        printf 'KeyLoad__ContainerImages__Server=%s\nKEYLOAD_IMAGE_RECEIPT=%s\n' "$KEYLOAD_SERVER_IMAGE" "$RUNNER_TEMP/keyload-images/image-receipt.json" >> "$GITHUB_ENV"
+    - name: Restore .NET packages
+      shell: bash
+      run: dotnet restore KeyLoad.slnx
+    - name: Build functional RF3 tests
+      shell: bash
+      run: dotnet build tests/KeyLoad.IntegrationTests --no-restore --configuration Release

.github/workflows/Features/BenchmarkComparisons/IsolatedCellSetup/AGENTS.md

@@ -82,12 +82,12 @@ runs:
         git -C control rev-parse HEAD > "$EVIDENCE_DIR/control-revision.txt"
         closure="scripts/Features/BenchmarkComparisons/site-isolated-dependencies.txt"
         cmp "control/$closure" "website/$closure"
-        [[ "$(wc -l < "control/$closure")" == 84 ]]
+        node control/scripts/Features/BenchmarkComparisons/session-source-closure.mjs --verify="control/$closure" > "$EVIDENCE_DIR/session-source-closure.json"
         LC_ALL=C sort -u "control/$closure" > "$EVIDENCE_DIR/isolated-dependencies.txt"
         cmp "control/$closure" "$EVIDENCE_DIR/isolated-dependencies.txt"
         sha256sum "website/$closure" >> "$EVIDENCE_DIR/executed-evidence-tools.sha256"
         while IFS= read -r relative; do
-          [[ "$relative" =~ ^(scripts/Features/BenchmarkComparisons/[^/]+\.mjs|site/Features/BenchmarkComparisons/[^/]+\.mjs|site/scripts/build\.mjs|benchmarks/KeyLoad.Comparisons/Features/BenchmarkComparisons/(isolated-contract|Documents/document-contract)\.json)$ ]]
+          [[ "$relative" =~ ^(scripts/Features/BenchmarkComparisons/[^/]+\.(mjs|ps1)|scripts/Features/CodeQuality/functional-coverage\.(shared|parse|native-merge\.trx)\.ps1|site/Features/BenchmarkComparisons/[^/]+\.mjs|site/scripts/build\.mjs)$ ]]
           for root in control website; do
             [[ -f "$root/$relative" && ! -L "$root/$relative" ]]
             [[ "$(realpath -e "$root/$relative")" == "$GITHUB_WORKSPACE/$root/$relative" ]]
@@ -125,7 +125,7 @@ runs:
         dotnet build src/KeyLoad.AppHost/KeyLoad.AppHost.csproj --no-restore --configuration Release
         dotnet build tests/KeyLoad.ComparisonTests/KeyLoad.ComparisonTests.csproj --no-restore --configuration Release
         node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/SiteQualificationStartupTests/*' --KeyLoadTests:ReportTrx=true "--KeyLoadTests:ResultsDirectory=$EVIDENCE_DIR/startup-tests"
-        node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/SiteOptionalBenchmarkSelectionTests/*' --KeyLoadTests:ReportTrx=true "--KeyLoadTests:ResultsDirectory=$EVIDENCE_DIR/selection-tests"
+        node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/SessionCohortContractTests/*' --KeyLoadTests:ReportTrx=true "--KeyLoadTests:ResultsDirectory=$EVIDENCE_DIR/selection-tests"
         dotnet build tests/KeyLoad.Analyzers.Tests/KeyLoad.Analyzers.Tests.csproj --no-restore --configuration Release
         mkdir -p "$EVIDENCE_DIR/analyzer-coverage"
         cp scripts/Features/CodeQuality/site-analyzer-coverage.settings.xml "$EVIDENCE_DIR/analyzer-coverage/coverage.config.xml"

.github/workflows/Features/BenchmarkComparisons/IsolatedCellTeardown/AGENTS.md

@@ -0,0 +1,9 @@
+# SessionContracts
+
+Owns the shared native benchmark preparation checks in `action.yml`. Read the root and workflow policies and ADR-142. Execute only inside Linux Benchmarks; preserve original image, topology, readiness, failure, cancellation and workload assertions. This composite never produces website measurements. Root integrates shared workflow changes.
+
+Every filter must name retained TUnit test classes. The comparison entry requires at least one discovered test; an empty filter is a failure, never a skipped qualification. Shared-session process admission is exercised by `ComparisonHostStartupTests`, `ComparisonHostCleanupTests` and `BenchmarkSessionNativeParserTests`. `BenchmarkSessionProgressTests` uses an explicitly controlled development target to check acknowledgement, cancellation and failed-oracle accounting; it is not native database performance evidence. `DocumentComparisonNativeTests` retains real SurrealDB/HelixDB shared-session correctness and cancellation against small functional corpora.
+
+The native resource-model group retains all current database topology, credential, private-volume and foreign-selection checks. Native shutdown tests retain original primary/cleanup failures and joined resource/log ownership. Keep MongoDB readiness, pinned-image inspection and KeyLoad resource/admission gates independent.
+
+The three removed `IsolatedTimeSeries*ResourceTests` classes belonged exclusively to retired TimeSeries benchmark dispatch. Do not retain their empty filters or substitute pgvector for TimescaleDB. `TimeSeriesIntensivePinnedImageTests`, `TimeSeriesWorkloadTests` and `TimeSeriesPackageVersionTests` remain active. Native Timescale family deployment, `TimeSeriesForeignSchemaCollision` and `TimescaleTimeSeriesIntensiveSchemaAtomicityRegression` remain pending their actual future family integration; pinned image facts do not qualify them.

.github/workflows/Features/BenchmarkComparisons/QualifySite/action.yml

@@ -0,0 +1,96 @@
+name: Native benchmark preparation contracts
+description: Run the complete native preparation checks once before database sessions.
+runs:
+  using: composite
+  steps:
+    - name: Check native KurrentDB discovery settings
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/IsolatedKurrentDiscoverySettingsTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/kurrent-discovery
+    - name: Check exact OpenSearch vector query contract
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/OpenSearchVectorQueryTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/opensearch-query
+    - name: Check exact OpenSearch vector response contract
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/OpenSearchVectorResponseTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/opensearch-response
+    - name: Check TimescaleDB Docker image and startup tools
+      env:
+        KEYLOAD_TIMESERIES_IMAGE_FACTS_DIRECTORY: ${{ runner.temp }}/keyload-timeseries-image-facts
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/TimeSeriesIntensivePinnedImageTests/*' --KeyLoadTests:ResultsDirectory=TestResults/timeseries-image
+    - name: Save TimescaleDB image checks
+      if: always()
+      uses: actions/upload-artifact@043fb46d1a93c77aae656e7c1c64a875d1fc6a0a
+      with:
+        name: timeseries-native-pinned-image-facts
+        path: |
+          ${{ runner.temp }}/keyload-timeseries-image-facts/**
+          **/TestResults/**
+          artifacts/code-quality/**
+        if-no-files-found: error
+        retention-days: 90
+    - name: Test database cluster resource models
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/Isolated*Resource*/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/resource-models
+    - name: Test native session shutdown and original failure retention
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/(IsolatedNativeTeardownSuccessTests|IsolatedNativeTeardownFailureTests|IsolatedNativeLogCaptureConcurrentDisposeTests)/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/session-shutdown --KeyLoadTests:ReportTrx=true
+    - name: Test MongoDB readiness configuration
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/MongoReadinessResourceTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/mongo-resource
+    - name: Test MongoDB node identity
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/MongoReadinessIdentityTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/mongo-identity
+    - name: Test MongoDB startup failure handling
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/MongoReadinessTaskFailureTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/mongo-task-failure
+    - name: Test KurrentDB storage failure handling
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/IsolatedKurrentVolumeRegressionFailureTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/kurrent-volume-failure
+    - name: Test KeyLoad request admission
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/IsolatedKeyLoadAdmissionTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/keyload-admission
+    - name: Test bounded database log buffers
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/ComparisonResourceLogBufferTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/log-buffer
+    - name: Test live benchmark progress capture
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/ComparisonProgress*/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/live-progress
+    - name: Test shared native session entry and safe failures
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/(ComparisonHostStartupTests|ComparisonHostCleanupTests|BenchmarkSessionNativeParserTests)/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/live-entry
+    - name: Test shared session live acknowledgement and cancellation counters
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/BenchmarkSessionProgressTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/live-counters
+    - name: Test document resource budgets and current native topology
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/(DocumentResourceBudgetTests|NativeTopologyAdmissionTests)/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/document-admission --KeyLoadTests:ReportTrx=true
+    - name: Test controlled native shared-session operations and cancellation
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/DocumentComparisonNativeTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/document-native --KeyLoadTests:ReportTrx=true
+    - name: Test KeyLoad replay admission
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/IsolatedKeyLoadReplayAdmissionTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/replay-admission
+    - name: Test Redis replica error reporting
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/RedisReplicaDiagnosticTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/redis-diagnostics
+    - name: Test MongoDB startup error reporting
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/MongoBootstrapDiagnosticTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/mongo-diagnostics
+    - name: Test KurrentDB cleanup error reporting
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/KurrentCleanupDiagnosticTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/kurrent-diagnostics
+    - name: Test KurrentDB stream ownership
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/KurrentStreamOwnershipTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/kurrent-stream-ownership
+    - name: Test database replay logs
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/ComparisonReplayDiagnosticLogTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/replay-log
+    - name: Test database replay log retention
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/ComparisonReplayDiagnosticRetentionTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/replay-retention
+    - name: Test time-series workload correctness
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/TimeSeriesWorkloadTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/timeseries-workload
+    - name: Check published time-series package version
+      shell: bash
+      run: node scripts/Features/TestInfrastructure/run-tests.mjs --KeyLoadTests:Suite=comparison '--KeyLoadTests:Filter=/*/*/TimeSeriesPackageVersionTests/*' --KeyLoadTests:ResultsDirectory=TestResults/comparison-images/timeseries-package

.github/workflows/Features/BenchmarkComparisons/SessionContracts/AGENTS.md

@@ -0,0 +1,3 @@
+# SessionSetup
+
+Owns setup for one native database/topology session under ADR-142. Follow root and workflow policies. Authenticate the same-run job, import original source-bound native images and the verified common compiled runtime, and preserve original import receipts. Never restore, build, reset a database, run a workload or manufacture results here. Linux GitHub execution only; scoped image registry shutdown belongs to the existing native lifetime.