Merge pull request #130 from redhat-performance/fix-sd-updater-multip…

…le-results

feat(SDupdater): Make SDupdater less strict about duplicate test results and just pick the oldest one

opl/status_data_updater.py

@@ -47,7 +47,7 @@ def get_session():
  return session
 
 
-def _es_get_test(session, args, key, val, size=1, sort_by="started"):
+def _es_get_test(session, args, key, val, size=1, sort_by="started", sort_order="desc"):
  url = f"{args.es_server}/{args.es_index}/_search"
  headers = {
  "Content-Type": "application/json",
@@ -60,7 +60,7 @@ def _es_get_test(session, args, key, val, size=1, sort_by="started"):
  },
  "sort": {
  sort_by: {
- "order": "desc",
+ "order": sort_order,
  },
  },
  "size": size,
@@ -338,18 +338,21 @@ def _get_es_dashboard_result_for_run_id(session, args, run_id, test=None):
  ["result_id.keyword", "test.keyword"],
  [run_id, test],
  sort_by="date",
+ sort_order="asc",
  )
  else:
  response = _es_get_test(
- session, args, ["result_id.keyword"], [run_id], sort_by="date"
+ session,
+ args,
+ ["result_id.keyword"],
+ [run_id],
+ sort_by="date",
+ sort_order="asc",
  )
  if response["hits"]["total"]["value"] == 0:
  return (None, None, None)
- else:
- assert (
- response["hits"]["total"]["value"] == 1
- ), "There have to be exactly one result"
  try:
+ # In combination with _es_get_test()'s sort_order="asc", get the oldest result.
  source = response["hits"]["hits"][0]
  except IndexError:
  logging.debug(f"Failed to find dashboard result in ES for {run_id}")