Merge pull request #273 from nulib/add-load-tests

Add Locust load tests and instructions

.tool-versions

@@ -1,4 +1,4 @@
 nodejs 20.15.0
 java corretto-19.0.1.10.1
 aws-sam-cli 1.107.0
-python 3.10.5
+python 3.12.2

loadtest/README.md

@@ -0,0 +1,24 @@
+# API Load Testing
+
+The API's load tests are written using the [Locust](https://locust.io/) load testing framework.
+
+## Usage
+
+### Set up dependencies
+```shell
+python -m venv ./.venv
+. ./.venv/bin/activate
+pip install -r requirements.txt
+```
+
+### Start server
+```shell
+API_BASE_URL=https://dcapi.rdc-staging.library.northwestern.edu/api/v2 # or whatever
+locust -f locustfile.py --host=${API_BASE_URL} --processes -1
+```
+
+### Run load tests
+1. Open http://localhost:8089/ in a browser.
+2. Customize test parameters (peak user count, ramp up, run time, etc.).
+3. Click **Start**.
+4. You can click around the UI while the test is running to see statistics, graphs, etc.

loadtest/locustfile.py

@@ -0,0 +1,73 @@
+from locust import HttpUser, task, between
+import random
+
+class DcApiUser(HttpUser):
+  wait_time = between(1, 5)
+
+  @task(1)
+  def page_through_works(self):
+    with self.client.rename_request("/search/works (paged)"):
+      response = self.client.get('/search/works').json()
+      for i in range(4):
+        next_url = response['pagination']['next_url']
+        response = self.client.get(next_url).json()
+
+  @task(1)
+  def search_works(self):
+    with self.client.rename_request("/search/works with query (paged)"):
+      query = { 'query': { 'term': { 'title': 'baez' } } }
+      response = self.client.post('/search/works', json=query).json()
+      for i in range(4):
+        next_url = response['pagination']['next_url']
+        response = self.client.get(next_url).json()
+
+  @task(3)
+  def load_collection_as_json(self):
+    id = random.choice(self.collection_ids)
+    self.client.get(f'/collections/{id}', name="/collections/:id")
+
+  @task(3)
+  def load_collection_as_iiif(self):
+    id = random.choice(self.collection_ids)
+    self.client.get(f'/collections/{id}?as=iiif', name="/collections/:id?as=iiif")
+
+  @task(3)
+  def load_work_as_json(self):
+    id = random.choice(self.work_ids)
+    self.client.get(f'/works/{id}', name="/works/:id")
+
+  @task(3)
+  def load_work_as_iiif(self):
+    id = random.choice(self.work_ids)
+    self.client.get(f'/works/{id}?as=iiif', name="/works/:id?as=iiif")
+
+  @task(3)
+  def load_work_thumbnail(self):
+    id = random.choice(self.work_ids)
+    self.client.get(f'/works/{id}/thumbnail', name="/works/:id/thumbnail")
+
+  @task(3)
+  def load_file_set(self):
+    id = random.choice(self.file_set_ids)
+    self.client.get(f'/file-sets/{id}', name="/file-sets/:id")
+
+  def on_start(self):
+    response = self.random_docs('works', include='id,file_sets.id')
+    self.work_ids = [doc['id'] for doc in response['data']]
+    self.file_set_ids = [file_set['id'] for item in response['data'] for file_set in item['file_sets']]
+    self.collection_ids = [doc['id'] for doc in self.random_docs('collections')['data']]
+
+  def random_docs(self, type, count=100, include='id'):
+    query = {
+      "size": count,
+      "query": {
+        "function_score": {
+          "query": { "match_all": {} },
+          "random_score": {}
+        }
+      }
+    }
+
+    response = self.client.post(f'/search/{type}?_source_includes={include}', json=query, name=f'Load {count} random {type}')
+    json = response.json()
+    return json

loadtest/requirements.txt

@@ -0,0 +1 @@
+locust==2.32.2