[Findit] Process analysis requests of Waterfall failures concurrently.

appengine/findit/handlers/process_failure_analysis_requests.py

 # Copyright 2016 The Chromium Authors. All rights reserved.
 # Use of this source code is governed by a BSD-style license that can be
 # found in the LICENSE file.
 
 import json
+import logging
 
 from common import constants
 from common.base_handler import BaseHandler
 from common.base_handler import Permission
 from common.http_client_appengine import HttpClientAppengine
 from waterfall import buildbot
 from waterfall import build_failure_analysis_pipelines
 from waterfall import build_util
 
 
 def _TriggerNewAnalysesOnDemand(builds):
   for build in builds:
     master_name = build['master_name']
     builder_name = build['builder_name']
     build_number = build['build_number']
     failed_steps = build.get('failed_steps')
 
-    # TODO(stgao): make builder_alerts send information of whether a build
+    # TODO(stgao): make alerts-dispatcher send information of whether a build
     # is completed.
     build = build_util.DownloadBuildData(
         master_name, builder_name, build_number)
     if not build or not build.data:
+      logging.error('Failed to retrieve build data for %s/%s/%s, steps=%s',
+                   master_name, builder_name, build_number, repr(failed_steps))
       continue  # Skip the build, wait for next request to recheck.
 
     build_info = buildbot.ExtractBuildInfo(
         master_name, builder_name, build_number, build.data)
 
     build_failure_analysis_pipelines.ScheduleAnalysisIfNeeded(
         master_name, builder_name, build_number, failed_steps=failed_steps,
         build_completed=build_info.completed,
         force=False, queue_name=constants.WATERFALL_ANALYSIS_QUEUE)
 
 
-class TriggerAnalyses(BaseHandler):
+class ProcessFailureAnalysisRequests(BaseHandler):
   """Triggers new analyses on demand.
 
   This handler checks the build failures in the request, and triggers new
-  analyes for a build in two situations:
+  analyses for a build in two situations:
   1. A new step failed.
   2. The build became completed after last analysis. This will potentially
      trigger a try-job run.
   """
 
   PERMISSION_LEVEL = Permission.ADMIN
 
   def HandlePost(self):
     builds = json.loads(self.request.body).get('builds', [])
     _TriggerNewAnalysesOnDemand(builds)