Nightly test fix for crawler 1.7.0 (#2789)

- Use latest crawler image for tests
- Due to webrecorder/browsertrix-crawler#861 change, a crawl with no
successful pages should be treated as failed. Update fixture to allow
both failed or complete state for backwards compatibility for now.

backend/test_nightly/conftest.py

@@ -306,7 +306,7 @@ def error_crawl_id(admin_auth_headers, default_org_id):
             headers=admin_auth_headers,
         )
         data = r.json()
-        if data["state"] == "complete":
+        if data["state"] in ("failed", "complete"):
             return crawl_id
         time.sleep(5)
 

chart/test/test.yaml

@@ -24,7 +24,7 @@ crawler_channels:
     image: "docker.io/webrecorder/browsertrix-crawler:latest"
 
   - id: test
-    image: "docker.io/webrecorder/browsertrix-crawler:1.7.0-beta.0"
+    image: "docker.io/webrecorder/browsertrix-crawler:latest"
 
 mongo_auth:
   # specify either username + password (for local mongo)