The following revision refers to this bug:
  https://chromium.googlesource.com/infra/infra/+/902c3c2a8a12331248d8d49e7515e5b7d6bd5a1e

commit 902c3c2a8a12331248d8d49e7515e5b7d6bd5a1e
Author: Jeffrey Li <lijeffrey@chromium.org>
Date: Fri Jun 08 20:31:07 2018

[Findit] Flake Analyzer - Adding monitoring for pipeline completion

In order to monitor real-time error rates, completion events and error
events need to be incremented near-simultaneously. Currently total
triggered pipelines are counted at triggered time, not completion time
which throws off the metric timings. e.g. If many analyses are triggered
at once but then relatively few analyses later, then after several hours
those analyses fail, the graph of triggered vs errors is actually incorrect
due to the time offset.

This change introduces a monitoring event when the main pipeline is
finalized also to trigger a completion event which should be used for
error rate metrics.

Bug:  850671 
Change-Id: I60b1dcca523e15a1f091730eab670e255f38ed8e
Reviewed-on: https://chromium-review.googlesource.com/1091491
Reviewed-by: Roberto Carrillo <robertocn@chromium.org>
Commit-Queue: Jeffrey Li <lijeffrey@chromium.org>

[modify] https://crrev.com/902c3c2a8a12331248d8d49e7515e5b7d6bd5a1e/appengine/findit/pipelines/flake_failure/analyze_flake_pipeline.py
[modify] https://crrev.com/902c3c2a8a12331248d8d49e7515e5b7d6bd5a1e/appengine/findit/common/monitoring.py