Add locust benchmarks

.github/workflows/test_backend.yml

@@ -37,9 +37,9 @@ jobs:
         with:
           poetry-version: "1.5.1"
       - run: poetry install
-      - run: poetry run black --check test_observer tests migrations scripts tasks
-      - run: poetry run ruff test_observer tests migrations scripts tasks
-      - run: poetry run mypy --explicit-package-bases test_observer tests migrations scripts tasks
+      - run: poetry run black --check --extend-exclude charm .
+      - run: poetry run ruff --extend-exclude charm .
+      - run: poetry run mypy --exclude charm --explicit-package-bases .
       - run: poetry run pytest
         env:
           TEST_DB_URL: postgresql+pg8000://postgres:password@localhost:5432/postgres

backend/README.md

@@ -116,3 +116,19 @@ DOCKER_BUILDKIT=1 docker build . -t your-choice-of-tag -f backend/Dockerfile.pro
 ```
 
 GitHub built images are also made available in the authorisation requiring private registry at `ghcr.io/canonical/test_observer/api` (every main branch commit and tag following the pattern `v*.*.*` gets stored there).
+
+## Loading a database dump
+
+If you have a database dump created using pg_dump, you can load it your local database instance in k8s. To do that simply run the following:
+
+```bash
+kubectl exec -i services/test-observer-db -- pg_restore -U postgres -d postgres --clean --if-exists --no-owner < path/to/db.dump
+```
+
+## Benchmarks
+
+We use [locust](https://locust.io) to benchmark our API. Our benchmarks live in `locustfile.py` and can be run independantly via:
+
+```bash
+locust
+```

backend/locustfile.py

@@ -0,0 +1,35 @@
+import random
+import time
+
+from locust import HttpUser, between, task
+
+
+class APIUser(HttpUser):
+    wait_time = between(1, 5)
+
+    @task
+    def normal_usage(self):
+        artefacts = self._get_artefacts()
+        time.sleep(1)
+        for artefact in random.choices(artefacts, k=10):
+            test_executions = self._get_test_executions(artefact)
+            time.sleep(1)
+            for te in random.choices(test_executions, k=30):
+                self._get_test_results(te)
+                time.sleep(1)
+
+    def _get_artefacts(self) -> list[dict]:
+        return self.client.get("/v1/artefacts").json()
+
+    def _get_test_executions(self, artefact: dict) -> list[dict]:
+        json = self.client.get(
+            f"/v1/artefacts/{artefact['id']}/builds",
+            name="/v1/artefacts/<id>/builds",
+        ).json()
+        return [te for build in json for te in build["test_executions"]]
+
+    def _get_test_results(self, test_execution: dict) -> list[dict]:
+        return self.client.get(
+            f"/v1/test-executions/{test_execution['id']}/test-results",
+            name="/v1/test-executions/<id>/test-results",
+        ).json()