Route content archive ingestion tasks to the edx_content queue

import_content_files (added in #2938), ingest_edx_run_archive, and the
canvas archive tasks were never added to task_routes, so their S3
listings and archive parsing ran on the default celery worker. During
mitxonline backpopulates this OOM-killed the worker and took
get_mitxonline_data down with it. Route them to the edx_content queue
alongside the rest of the content-archive tasks.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Author: mbertrand

main/celery.py

@@ -37,4 +37,8 @@
     "learning_resources_search.tasks.deindex_run_content_files": {
         "queue": "edx_content"
     },
+    "learning_resources.tasks.import_content_files": {"queue": "edx_content"},
+    "learning_resources.tasks.ingest_edx_run_archive": {"queue": "edx_content"},
+    "learning_resources.tasks.ingest_canvas_course": {"queue": "edx_content"},
+    "learning_resources.tasks.sync_canvas_courses": {"queue": "edx_content"},
 }