diff --git a/.github/workflows/docker-tests.yml b/.github/workflows/docker-tests.yml
index fdf2b1a5238f..5180dacf7b5c 100644
--- a/.github/workflows/docker-tests.yml
+++ b/.github/workflows/docker-tests.yml
@@ -16,6 +16,12 @@ on:
 
 jobs:
   test_docker_image:
+    strategy:
+      # we could run all the flavors in a single pytest job,
+      # but let's split them (and the image build) in multiple runners for performance
+      matrix:
+        flavor: ["slim", "debian"]
+
     runs-on: ubuntu-latest
     timeout-minutes: 30
     steps:
@@ -30,37 +36,13 @@ jobs:
         with:
           config: .github/buildkit.toml
 
-      - name: "Build flavor: slim"
-        uses: docker/build-push-action@v5
-        with:
-          context: .
-          file: ./Dockerfile
-          tags: crowdsecurity/crowdsec:test-slim
-          target: slim
-          platforms: linux/amd64
-          load: true
-          cache-from: type=gha
-          cache-to: type=gha,mode=min
-
-      - name: "Build flavor: full"
-        uses: docker/build-push-action@v5
-        with:
-          context: .
-          file: ./Dockerfile
-          tags: crowdsecurity/crowdsec:test
-          target: full
-          platforms: linux/amd64
-          load: true
-          cache-from: type=gha
-          cache-to: type=gha,mode=min
-
-      - name: "Build flavor: full (debian)"
+      - name: "Build image"
         uses: docker/build-push-action@v5
         with:
           context: .
-          file: ./Dockerfile.debian
-          tags: crowdsecurity/crowdsec:test-debian
-          target: full
+          file: ./Dockerfile${{ matrix.flavor == 'debian' && '.debian' || '' }}
+          tags: crowdsecurity/crowdsec:test${{ matrix.flavor == 'full' && '' || join('-', matrix.flavor) }}
+          target: ${{ matrix.flavor == 'debian' && 'full' || matrix.flavor }}
           platforms: linux/amd64
           load: true
           cache-from: type=gha
@@ -95,9 +77,10 @@ jobs:
       - name: "Run tests"
         env:
           CROWDSEC_TEST_VERSION: test
-          CROWDSEC_TEST_FLAVORS: slim,debian
+          CROWDSEC_TEST_FLAVORS: ${{ matrix.flavor }}
           CROWDSEC_TEST_NETWORK: net-test
           CROWDSEC_TEST_TIMEOUT: 90
+        # running serially to reduce test flakiness
         run: |
           cd docker/test
-          pipenv run pytest -n 2 --durations=0 --color=yes
+          pipenv run pytest -n 1 --durations=0 --color=yes