Add Redis 8.x/Valkey CI test matrix and optional Valkey loadtests (#47180)

.github/workflows/loadtest-infra.yml

@@ -47,6 +47,24 @@ on:
         type: string
         default: 3
         required: true
+      fleet_redis_engine:
+        description: "Elasticache engine to deploy"
+        type: choice
+        options:
+        - redis
+        - valkey
+        default: redis
+        required: false
+      fleet_redis_engine_version:
+        description: "Elasticache engine version (must match the engine, e.g. 7.1 for redis, 8.0 for valkey)"
+        type: string
+        default: "7.1"
+        required: false
+      fleet_redis_parameter_group_family:
+        description: "Elasticache parameter group family (must match engine+version, e.g. redis7, valkey7, valkey8)"
+        type: string
+        default: "redis7"
+        required: false
       terraform_action:
         description: Dry run only? No "terraform apply"
         type: choice
@@ -79,6 +97,9 @@ env:
   TF_VAR_database_instance_count: "${{ inputs.fleet_database_instance_count }}"
   TF_VAR_redis_instance_size: "${{ inputs.fleet_redis_instance_size }}"
   TF_VAR_redis_instance_count: "${{ inputs.fleet_redis_instance_count }}"
+  TF_VAR_redis_engine: "${{ inputs.fleet_redis_engine }}"
+  TF_VAR_redis_engine_version: "${{ inputs.fleet_redis_engine_version }}"
+  TF_VAR_redis_parameter_group_family: "${{ inputs.fleet_redis_parameter_group_family }}"
 
 permissions:
   id-token: write

.github/workflows/test-go-suite.yaml

@@ -14,6 +14,11 @@ on:
         required: false
         type: string
         default: ''
+      redis:
+        description: 'Redis/Valkey Docker image (e.g., "redis:8.0", "valkey/valkey:8.0"). Leave empty to use the docker-compose default (redis:6.2).'
+        required: false
+        type: string
+        default: ''
       cover_pkg:
         description: 'Go coverage package pattern (e.g., "github.com/fleetdm/fleet/v4/server/activity/..."). Defaults to all fleet packages.'
         required: false
@@ -101,12 +106,19 @@ jobs:
 
     - name: Compute artifact prefix
       run: |
+        PREFIX="${{ inputs.suite }}"
         if [[ -n "${{ inputs.mysql }}" ]]; then
           MYSQL_ID=$(echo "${{ inputs.mysql }}" | tr -d ':')
-          echo "ARTIFACT_PREFIX=${{ inputs.suite }}-${MYSQL_ID}" >> $GITHUB_ENV
-        else
-          echo "ARTIFACT_PREFIX=${{ inputs.suite }}" >> $GITHUB_ENV
+          PREFIX="${PREFIX}-${MYSQL_ID}"
         fi
+        # Append the Redis/Valkey id only when an explicit image is requested, so the
+        # default callers keep their existing artifact names and matrix legs that vary
+        # only by Redis image don't collide on artifact upload.
+        if [[ -n "${{ inputs.redis }}" ]]; then
+          REDIS_ID=$(echo "${{ inputs.redis }}" | tr -d ':/')
+          PREFIX="${PREFIX}-${REDIS_ID}"
+        fi
+        echo "ARTIFACT_PREFIX=${PREFIX}" >> $GITHUB_ENV
 
     - name: Set Go race setting on schedule
       if: ${{ inputs.is_cron }}
@@ -131,7 +143,7 @@ jobs:
     - name: Start Infra Dependencies
       if: ${{ env.NEED_DOCKER }}
       # Use & to background this
-      run: FLEET_MYSQL_IMAGE=${{ inputs.mysql }} $DOCKER_COMMAND &
+      run: FLEET_REDIS_IMAGE=${{ inputs.redis }} FLEET_MYSQL_IMAGE=${{ inputs.mysql }} $DOCKER_COMMAND &
 
     - name: Add TLS certificate for SMTP Tests
       if: ${{ env.NEED_DOCKER }}
@@ -199,7 +211,7 @@ jobs:
           docker compose down
 
           echo "Restarting containers..."
-          FLEET_MYSQL_IMAGE=${{ inputs.mysql }} $DOCKER_COMMAND &
+          FLEET_REDIS_IMAGE=${{ inputs.redis }} FLEET_MYSQL_IMAGE=${{ inputs.mysql }} $DOCKER_COMMAND &
 
           # Give containers a moment to start
           sleep 10

.github/workflows/test-go.yaml

@@ -105,6 +105,30 @@ jobs:
       FLEET_RELEASE_GITHUB_PAT: ${{ secrets.FLEET_RELEASE_GITHUB_PAT }}
       SLACK_G_HELP_ENGINEERING_WEBHOOK_URL: ${{ secrets.SLACK_G_HELP_ENGINEERING_WEBHOOK_URL }}
 
+  # ──────────────────────────────────────────────────────────────────────────
+  # Extended Redis/Valkey coverage: only on the nightly cron schedule.
+  # Runs the Redis-touching suites against newer Redis versions and Valkey, in
+  # both standalone and cluster mode. MySQL is pinned to the always-run version
+  # (single DB dimension, not NxM). Redis 6.2 is already covered by the always-run
+  # test-go job above, so it's omitted here.
+  # make sure to update supported versions docs when Redis/Valkey versions change
+  # ──────────────────────────────────────────────────────────────────────────
+  test-go-extended-redis:
+    if: github.event_name == 'schedule'
+    strategy:
+      matrix:
+        suite: ["integration-core", "integration-enterprise", "integration-mdm", "main", "service"]
+        redis: ["redis:7.4", "redis:8.0", "valkey/valkey:7.2", "valkey/valkey:8.0"]
+    uses: ./.github/workflows/test-go-suite.yaml
+    with:
+      suite: ${{ matrix.suite }}
+      mysql: "mysql:8.0.44"
+      redis: ${{ matrix.redis }}
+      is_cron: true
+    secrets:
+      FLEET_RELEASE_GITHUB_PAT: ${{ secrets.FLEET_RELEASE_GITHUB_PAT }}
+      SLACK_G_HELP_ENGINEERING_WEBHOOK_URL: ${{ secrets.SLACK_G_HELP_ENGINEERING_WEBHOOK_URL }}
+
   # Based on https://github.com/micromdm/nanomdm/blob/main/.github/workflows/on-push-pr.yml#L87
   test-go-nanomdm:
     runs-on: 'ubuntu-latest'
@@ -224,7 +248,7 @@ jobs:
 
   # We upload all backend coverage in one step so that we're less like to end up in a situation with a partial coverage report.
   upload-coverage:
-    needs: [test-go-no-db, test-go, test-go-extended-mysql, test-go-nanomdm]
+    needs: [test-go-no-db, test-go, test-go-extended-mysql, test-go-extended-redis, test-go-nanomdm]
     # Run even if extended-mysql was skipped (non-cron) or individual suites failed.
     if: always()
     runs-on: ubuntu-latest
@@ -251,7 +275,7 @@ jobs:
   # Our Go test suites are run with continue-on-error: true, so they don't contribute to the workflow pass/fail.
   # This job explicitly checks if any Go test suites have failed and marks the overall workflow with the proper pass/fail status.
   aggregate-result:
-    needs: [test-go-no-db, test-go, test-go-extended-mysql]
+    needs: [test-go-no-db, test-go, test-go-extended-mysql, test-go-extended-redis]
     # Run even if extended-mysql was skipped (non-cron) or individual suites failed.
     if: always()
     runs-on: ubuntu-latest