[Infra] Remove CCI/GHA test duplication and semantically shard proxy DB tests

.circleci/config.yml

@@ -563,39 +563,6 @@ jobs:
           paths:
             - realtime_translation_coverage.xml
             - realtime_translation_coverage
-  mcp_testing:
-    docker:
-      - *python312_image
-    working_directory: ~/project
-
-    steps:
-      - checkout
-      - setup_google_dns
-      - install_uv
-      - run:
-          name: Install Dependencies
-          command: |
-            uv sync --frozen --all-groups --all-extras --python 3.12
-      # Run pytest and generate JUnit XML report
-      - run:
-          name: Run tests
-          command: |
-            uv run --no-sync python -m pytest -vv tests/mcp_tests --cov=litellm --cov-report=xml -x -s -v --junitxml=test-results/junit.xml --durations=5 -n 2
-          no_output_timeout: 15m
-      - run:
-          name: Rename the coverage files
-          command: |
-            mv coverage.xml mcp_coverage.xml
-            mv .coverage mcp_coverage
-
-      # Store test results
-      - store_test_results:
-          path: test-results
-      - persist_to_workspace:
-          root: .
-          paths:
-            - mcp_coverage.xml
-            - mcp_coverage
   agent_testing:
     docker:
       - *python312_image
@@ -794,39 +761,6 @@ jobs:
           paths:
             - search_coverage.xml
             - search_coverage
-  # Split litellm_mapped_tests into parallel jobs
-  litellm_mapped_tests_proxy_part1:
-    docker:
-      - *python312_image
-    working_directory: ~/project
-    resource_class: large
-    steps:
-      - setup_litellm_test_deps
-      - run:
-          name: Run proxy tests part 1 (high-volume directories)
-          command: |
-            uv run --no-sync python -m prisma generate
-            export PYTHONUNBUFFERED=1
-            uv run --no-sync python -m pytest tests/test_litellm/proxy/guardrails tests/test_litellm/proxy/management_endpoints tests/test_litellm/proxy/_experimental tests/test_litellm/proxy/client tests/test_litellm/proxy/auth --junitxml=test-results/junit-proxy-part1.xml --durations=10 -n 4 --maxfail=5 --timeout=60 -vv --log-cli-level=WARNING -r A
-          no_output_timeout: 15m
-      - store_test_results:
-          path: test-results
-  litellm_mapped_tests_proxy_part2:
-    docker:
-      - *python312_image
-    working_directory: ~/project
-    resource_class: large
-    steps:
-      - setup_litellm_test_deps
-      - run:
-          name: Run proxy tests part 2 (all other tests)
-          command: |
-            uv run --no-sync python -m prisma generate
-            export PYTHONUNBUFFERED=1
-            uv run --no-sync python -m pytest tests/test_litellm/proxy --ignore=tests/test_litellm/proxy/guardrails --ignore=tests/test_litellm/proxy/management_endpoints --ignore=tests/test_litellm/proxy/_experimental --ignore=tests/test_litellm/proxy/client --ignore=tests/test_litellm/proxy/auth --junitxml=test-results/junit-proxy-part2.xml --durations=10 -n 4 --maxfail=5 --timeout=120 -vv --log-cli-level=WARNING -r A
-          no_output_timeout: 15m
-      - store_test_results:
-          path: test-results
   litellm_mapped_enterprise_tests:
     docker:
       - *python312_image
@@ -2072,7 +2006,7 @@ jobs:
       - run:
           name: Combine Coverage
           command: |
-            uv tool run --from 'coverage[toml]==7.10.6' coverage combine realtime_translation_coverage ocr_coverage search_coverage mcp_coverage logging_coverage audio_coverage local_testing_part1_coverage local_testing_part2_coverage pass_through_unit_tests_coverage batches_coverage guardrails_coverage redis_caching_coverage
+            uv tool run --from 'coverage[toml]==7.10.6' coverage combine realtime_translation_coverage ocr_coverage search_coverage logging_coverage audio_coverage local_testing_part1_coverage local_testing_part2_coverage pass_through_unit_tests_coverage batches_coverage guardrails_coverage redis_caching_coverage
             uv tool run --from 'coverage[toml]==7.10.6' coverage xml
       - codecov/upload:
           file: ./coverage.xml
@@ -2407,8 +2341,6 @@ workflows:
           filters: *main_branches
       - realtime_translation_testing:
           filters: *main_branches
-      - mcp_testing:
-          filters: *main_branches
       - agent_testing:
           filters: *main_branches
       - guardrails_testing:
@@ -2423,10 +2355,6 @@ workflows:
           filters: *main_branches
       - litellm_mapped_enterprise_tests:
           filters: *main_branches
-      - litellm_mapped_tests_proxy_part1:
-          filters: *main_branches
-      - litellm_mapped_tests_proxy_part2:
-          filters: *main_branches
       - batches_testing:
           filters: *main_branches
       - litellm_utils_testing:
@@ -2444,14 +2372,11 @@ workflows:
       - upload-coverage:
           requires:
             - realtime_translation_testing
-            - mcp_testing
             - agent_testing
             - google_generate_content_endpoint_testing
             - guardrails_testing
             - ocr_testing
             - search_testing
-            - litellm_mapped_tests_proxy_part1
-            - litellm_mapped_tests_proxy_part2
             - litellm_mapped_enterprise_tests
             - batches_testing
             - litellm_utils_testing

.github/workflows/_test-unit-services-base.yml

@@ -32,41 +32,39 @@ on:
         required: false
         type: boolean
         default: false
+      dist:
+        description: "pytest-xdist distribution mode (loadscope|load|worksteal|loadfile|no)"
+        required: false
+        type: string
+        default: "loadscope"
       artifact-name:
         description: "Unique name for the coverage artifact (must be unique per run)"
         required: false
         type: string
         default: "run"
-    secrets:
-      DATABASE_URL:
-        required: false
-      POSTGRES_USER:
-        required: false
-      POSTGRES_PASSWORD:
-        required: false
 
 permissions:
   contents: read
 
+# The postgres service container below is spawned per-job on localhost and
+# destroyed with the job. Nothing outside the runner can reach it. The
+# user/password/database here are not secrets — they're bootstrap values
+# for a throwaway container — so we hardcode them instead of attaching
+# every matrix shard to a GHA environment just to read three "secrets"
+# (which also produces a "temporarily deployed to …" notification on the
+# PR timeline per shard per push).
 jobs:
   run:
     name: Run tests
     runs-on: ubuntu-latest
     timeout-minutes: ${{ inputs.timeout-minutes }}
-    # Environment is derived from the enable-* flags, not caller-controllable.
-    # This prevents callers from passing arbitrary environment names to bypass secret scoping.
-    environment: >-
-      ${{
-        inputs.enable-postgres && 'integration-postgres' ||
-        ''
-      }}
 
     services:
       postgres:
         image: postgres@sha256:705a5d5b5836f3fcba0d02c4d281e6a7dd9ed2dd4078640f08a1e1e9896e097d # postgres:14
         env:
-          POSTGRES_USER: ${{ secrets.POSTGRES_USER }}
-          POSTGRES_PASSWORD: ${{ secrets.POSTGRES_PASSWORD }}
+          POSTGRES_USER: litellm
+          POSTGRES_PASSWORD: litellm
           POSTGRES_DB: litellm_test
         ports:
           - 5432:5432
@@ -114,7 +112,7 @@ jobs:
       - name: Run Prisma migrations
         if: ${{ inputs.enable-postgres }}
         env:
-          DATABASE_URL: ${{ secrets.DATABASE_URL }}
+          DATABASE_URL: "postgresql://litellm:litellm@localhost:5432/litellm_test"
         run: |
           uv run --no-sync prisma db push --schema litellm/proxy/schema.prisma --accept-data-loss
 
@@ -124,7 +122,8 @@ jobs:
           MAX_FAILURES: ${{ inputs.max-failures }}
           WORKERS: ${{ inputs.workers }}
           RERUNS: ${{ inputs.reruns }}
-          DATABASE_URL: ${{ inputs.enable-postgres && secrets.DATABASE_URL || '' }}
+          DIST: ${{ inputs.dist }}
+          DATABASE_URL: ${{ inputs.enable-postgres && 'postgresql://litellm:litellm@localhost:5432/litellm_test' || '' }}
         run: |
           if [ "${WORKERS}" = "0" ]; then
             uv run --no-sync pytest ${TEST_PATH:?} \
@@ -143,7 +142,7 @@ jobs:
               -n "${WORKERS}" \
               --reruns "${RERUNS}" \
               --reruns-delay 1 \
-              --dist=loadscope \
+              --dist="${DIST}" \
               --durations=20 \
               --cov=litellm \
               --cov-report=xml:coverage.xml \