Merge branch 'feat/add-pydantic-handling-decorator-for-req-res' into …

…refactor/apply-decorator

.github/PULL_REQUEST_TEMPLATE.md

@@ -1,3 +1,6 @@
+resolves #NNN (BA-MMM)
+<!-- replace NNN, MMM with the GitHub issue number and the corresponding Jira issue number. -->
+
 <!--
 Please precisely, concisely, and concretely describe what this PR changes, the rationale behind codes,
 and how it affects the users and other developers.

.github/labeler.yml

@@ -7,8 +7,6 @@ area:docs:
 - changed-files:
   - any-glob-to-any-file:
     - 'docs/**/*'
-    - 'src/ai/backend/manager/api/**/*.py'
-    - 'src/ai/backend/manager/models/gql.py'
 
 comp:storage-proxy:
 - changed-files:

.github/workflows/ci.yml

@@ -364,14 +364,20 @@ jobs:
         gha-cache-key: pants-cache-main-1-deploy-py${{ env.PROJECT_PYTHON_VERSION }}-${{ runner.os }}-${{ runner.arch }}
         named-caches-hash: ${{ hashFiles('python*.lock', 'tools/*.lock') }}
         cache-lmdb-store: 'false'
-    - name: Build fat packages
+    - name: Build both lazy and fat packages
       run: |
         pants --tag="scie" package '::'
-        # 'pants run' does not support parallelization
+        # Rename artifacts to have the platform suffix
+        platform_suffix="$(python scripts/get-platform-suffix.py)"
+        for artifact in dist/backendai-*; do
+          mv "${artifact}" "${artifact}-${platform_suffix}"
+        done
+        # Generate checksums. NOTE: 'pants run' does not support parallelization
         pants list --filter-tag-regex='checksum' '::' | xargs -n 1 pants run
-    - name: Clean up intermediate pex files
-      run: |
-        rm -rf dist/src.*/
+        # Merge checksums into a single file
+        cat dist/*.sha256 > dist/checksum.txt
+        sort -u dist/checksum.txt -o dist/checksum.txt
+        rm dist/*.sha256
     - name: Upload scies
       uses: actions/upload-artifact@v4
       with:

.github/workflows/update-api-schema.yml

@@ -4,9 +4,10 @@ on:
   pull_request:
     paths:
       - 'src/ai/backend/manager/models/**'
+      - 'src/ai/backend/manager/api/**'
 
 jobs:
-  graphql-updated:
+  api-updated:
     runs-on: ubuntu-latest
     permissions:
       contents: write
@@ -52,27 +53,28 @@ jobs:
         cache-lmdb-store: 'true'
     - name: Pants export
       run: pants export --resolve=python-default
-    - name: Create GraphQL schema dump
+    - name: Create api schema dump
       run: |
-        ./backend.ai mgr api dump-gql-schema --output src/ai/backend/manager/api/schema.graphql
+        ./backend.ai mgr api dump-gql-schema --output docs/manager/graphql-reference/schema.graphql
+        ./backend.ai mgr api dump-openapi --output docs/manager/rest-reference/openapi.json
     - name: Make commit message for changing change log file
       run: |
-        git add src/ai/backend/manager/api/schema.graphql
+        git add docs/manager/graphql-reference/schema.graphql
+        git add docs/manager/rest-reference/openapi.json
         author_name=$(git show -q --pretty='format:%an')
         author_email=$(git show -q --pretty='format:%ae')
         git config user.email "$author_email"
         git config user.name "$author_name"
         if [ -n "$(git diff --staged)" ]; then
           git commit \
-            -m "chore: update GraphQL schema dump" \
+            -m "chore: update api schema dump" \
             --author="$author_name <$author_email>" \
             --trailer "Co-authored-by: octodog <[email protected]>"
           git push
         fi
 
   graphql-inspector:
-    needs: graphql-updated
-    name: Check Schema
+    needs: api-updated
     runs-on: ubuntu-latest
     permissions:
       contents: read
@@ -82,6 +84,6 @@ jobs:
       - uses: actions/checkout@v4
       - uses: kamilkisiela/graphql-inspector@release-1717403590269
         with:
-          schema: '${{ github.base_ref }}:src/ai/backend/manager/api/schema.graphql'
+          schema: '${{ github.base_ref }}:docs/manager/graphql-reference/schema.graphql'
           rules: |
             gql-inspector-checker.js

.gitignore

@@ -42,6 +42,7 @@ coverage.xml
 .hypothesis/
 .pytest_cache/
 .mypy_cache/
+.ruff_cache/
 
 # Translations
 *.mo
@@ -78,7 +79,7 @@ venv/
 ENV/
 
 # IDE/vim
-.idea
+.idea/
 .swp
 .swo
 .*.swp
@@ -90,7 +91,7 @@ ENV/
 /.exrc
 .ripgreprc
 .neoconf.json
-/.zed
+.zed/
 /.ropeproject
 pyrightconfig.json
 
@@ -103,13 +104,13 @@ pyrightconfig.json
 /tools/pants-src
 
 # Local configurations
-/account-manager.toml
-/manager.toml
-/agent.toml
-/storage-proxy.toml
-/webserver.conf
-/cuda-mock.toml
-/mock-accelerator.toml
+/account-manager*.toml
+/manager*.toml
+/agent*.toml
+/storage-proxy*.toml
+/webserver*.conf
+/cuda-mock*.toml
+/mock-accelerator*.toml
 /.pants.env
 /.pants.bootstrap
 /.pants.rc

.readthedocs.yaml

@@ -16,8 +16,6 @@ build:
           exit 183;
         fi
     pre_build:
-      # Auto-generate REST API reference
-      - PYTHONPATH="src" python -m ai.backend.manager.openapi -o docs/manager/rest-reference/openapi.json
       - |
         if [ "$READTHEDOCS_VERSION_TYPE" != "external" ];
         then

BUILD

@@ -10,6 +10,7 @@ python_requirements(
         "python-dateutil": ["dateutil", "dateutil.parser", "dateutil.tz"],
         "python-json-logger": ["pythonjsonlogger"],
         "pyhumps": ["humps"],
+        "pyroscope-io": ["pyroscope"],
         "pyzmq": ["zmq"],
         "PyYAML": ["yaml"],
         "typing-extensions": ["typing_extensions"],

CHANGELOG.md

@@ -16,6 +16,45 @@ Changes
 
 <!-- towncrier release notes start -->
 
+## 25.1.1 (2025-01-20)
+No significant changes.
+
+
+## 25.1.0 (2025-01-20)
+
+### Features
+* Implement fine-grained seccomp profile managed by Backend.AI Agent. ([#3019](https://github.com/lablup/backend.ai/issues/3019))
+* Enable image rescanning by project. ([#3237](https://github.com/lablup/backend.ai/issues/3237))
+* Support auto-scaling of model services by observing proxy and app-specific metrics as configured by autoscaling rules bound to each endpoint ([#3277](https://github.com/lablup/backend.ai/issues/3277))
+* Deprecate the JWT-based `X-BackendAI-SSO` header to reduce complexity in authentication process for the pipeline service ([#3353](https://github.com/lablup/backend.ai/issues/3353))
+* Add Grafana and Prometheus to Docker Compose ([#3458](https://github.com/lablup/backend.ai/issues/3458))
+* Integrate Pyroscope with Backend.AI ([#3459](https://github.com/lablup/backend.ai/issues/3459))
+* Update SDK to retrieve and use IDs for VFolder API operations instead of names ([#3471](https://github.com/lablup/backend.ai/issues/3471))
+
+### Fixes
+* Refactor container registries' projects traversal logic of the image rescanning. ([#2979](https://github.com/lablup/backend.ai/issues/2979))
+* Fix regression of outdated `vfolder` GQL resolver. ([#3047](https://github.com/lablup/backend.ai/issues/3047))
+* Fix image without metadata label not working ([#3341](https://github.com/lablup/backend.ai/issues/3341))
+* Enforce VFolder name length restriction through the API schema, not by the DB column constraint ([#3363](https://github.com/lablup/backend.ai/issues/3363))
+* Fix password based SSH login not working on sessions based on certain images ([#3387](https://github.com/lablup/backend.ai/issues/3387))
+* Fix purge API to allow deletion of owner-deleted VFolders by directly retrieving VFolders using the folder ID ([#3388](https://github.com/lablup/backend.ai/issues/3388))
+* Fix certain customized images not being pushed to registry properly ([#3391](https://github.com/lablup/backend.ai/issues/3391))
+* Fix formatting errors when logging exceptions raised from the current local process that did not pass our custom serialization step ([#3410](https://github.com/lablup/backend.ai/issues/3410))
+* Fix scanning and loading container images with no labels at all (`null` in the image manifests) ([#3411](https://github.com/lablup/backend.ai/issues/3411))
+* Fix missing CPU architecture name lookup in `LocalRegistry` to directly scan and load container images from the local Docker daemon in dev setups ([#3420](https://github.com/lablup/backend.ai/issues/3420))
+* Utilization idle checker computes kernel resource usages correctly ([#3442](https://github.com/lablup/backend.ai/issues/3442))
+* Filter vfolders by status before initiating a vfolder deletion task ([#3446](https://github.com/lablup/backend.ai/issues/3446))
+* Fix a mis-implementation that has prevented using UUIDs to indicate an exact vfolder when invoking the vfolder REST API ([#3451](https://github.com/lablup/backend.ai/issues/3451))
+* Fix the required state output logic in the openopi reference documentation correctly ([#3460](https://github.com/lablup/backend.ai/issues/3460))
+* Raise exception if multiple VFolders exist in decorator ([#3465](https://github.com/lablup/backend.ai/issues/3465))
+
+### Documentation Updates
+* Deprecate non relay container registry GQL explicitly. ([#3231](https://github.com/lablup/backend.ai/issues/3231))
+
+### Miscellaneous
+* Upgrade pantsbuild from 2.21 to 2.23, replacing the scie plugin with the intrinsic pex's scie build support ([#3377](https://github.com/lablup/backend.ai/issues/3377))
+
+
 ## 24.12.1 (2025-01-04)
 
 ### Fixes

VERSION

@@ -1 +1 @@
-24.12.1
+25.1.1

changes/3511.feature.md

@@ -0,0 +1 @@
+Add new Pydantic handling api decorator for Request/Response validation

configs/account-manager/halfstack.toml

@@ -28,6 +28,13 @@ ssl-enabled = false
 pid-file = "./account-manager.pid"             # env: BACKEND_PID_FILE
 
 
+[pyroscope]
+enabled = true
+app_name = "backendai-half-account-manager"
+server_addr = "http://localhost:4040"
+sample-rate = 100
+
+
 [logging]
 level = "INFO"
 drivers = ["console"]

configs/account-manager/sample.toml

@@ -65,6 +65,12 @@ ssl-enabled = false
 pid-file = "./account-manager.pid"             # env: BACKEND_PID_FILE
 
 
+[pyroscope]
+enabled = true
+app-name = "backendai-half-account-manager"
+server-addr = "http://localhost:4040"
+sample-rate = 100
+
 [logging]
 level = "INFO"
 drivers = ["console"]

configs/agent/ci.toml

@@ -7,6 +7,11 @@ password = ""
 
 [agent]
 rpc-listen-addr = { host = "127.0.0.1", port = 6001 }
+# metric API service address
+service-addr = { host = "0.0.0.0", port = 6003 }
+ssl-enabled = false
+#ssl-cert = ""
+#ssl-key = ""
 agent-sock-port = 6007
 id = "i-travis"
 scaling-group = "default"
@@ -34,6 +39,13 @@ target-service = "backendai-agent.service"
 soft-reset-available = false
 
 
+[pyroscope]
+enabled = false
+# app-name = "backendai-half-agent"
+# server-addr = "http://localhost:4040"
+# sample-rate = 100
+
+
 [logging]
 level = "INFO"
 drivers = ["console"]

configs/agent/halfstack.toml

@@ -8,6 +8,11 @@ password = ""
 [agent]
 mode = "docker"
 rpc-listen-addr = { host = "127.0.0.1", port = 6001 }
+# metric API service address
+service-addr = { host = "0.0.0.0", port = 6003 }
+ssl-enabled = false
+#ssl-cert = ""
+#ssl-key = ""
 agent-sock-port = 6007
 # id = "i-something-special"
 scaling-group = "default"
@@ -42,6 +47,13 @@ target-service = "backendai-agent.service"
 soft-reset-available = false
 
 
+[pyroscope]
+enabled = true
+app-name = "backendai-half-agent"
+server-addr = "http://localhost:4040"
+sample-rate = 100
+
+
 [logging]
 level = "INFO"
 drivers = ["console"]

configs/agent/sample.toml

@@ -216,6 +216,13 @@ target-service = "backendai-agent.service"
 soft-reset-available = false
 
 
+[pyroscope]
+enabled = true
+app-name = "backendai-half-agent"
+server-addr = "http://localhost:4040"
+sample-rate = 100
+
+
 [logging]
 # One of: "NOTSET", "DEBUG", "INFO", "WARNING", "ERROR", "CRITICAL"
 # Set the global logging level.