TypeError: unsupported operand type(s) for *: 'float' and 'NoneType'
at Merge (/base/data/home/apps/s~findit-for-me/waterfall-backend:10909-fba6c55.402923779948242132/model/flake/master_flake_analysis.py:134)
at AppendOrMergeDataPoint (/base/data/home/apps/s~findit-for-me/waterfall-backend:10909-fba6c55.402923779948242132/model/flake/master_flake_analysis.py:354)
at run (/base/data/home/apps/s~findit-for-me/waterfall-backend:10909-fba6c55.402923779948242132/waterfall/flake/update_flake_analysis_data_points_pipeline.py:114)
at _run_internal (/base/data/home/apps/s~findit-for-me/waterfall-backend:10909-fba6c55.402923779948242132/first_party/pipeline_utils/appengine_third_party_pipeline_python_src_pipeline/pipeline.py:1113)
at evaluate (/base/data/home/apps/s~findit-for-me/waterfall-backend:10909-fba6c55.402923779948242132/first_party/pipeline_utils/appengine_third_party_pipeline_python_src_pipeline/pipeline.py:2158)

The root cause is a flake swarming task ended in error, causing a data point with None as iterations to be saved. When trying to merge a new data point, None was multipled by the passrate leading to this error.

We should not save data points for tasks that ended in error.

Tasks that end in error should be retried.