[beam] branch aaltay-patch-1 created (now 9b043c9)

[al...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

altay pushed a change to branch aaltay-patch-1
in repository https://gitbox.apache.org/repos/asf/beam.git.


      at 9b043c9  Increase overhaed budget for test_sampler_transition_overhead

This branch includes the following new commits:

     new 9b043c9  Increase overhaed budget for test_sampler_transition_overhead

The 1 revisions listed above as "new" are entirely new to this
repository and will be described in separate emails.  The revisions
listed as "add" were already present in the repository and have only
been added to this reference.

[beam] 01/01: Increase overhaed budget for test_sampler_transition_overhead

[al...@apache.org]

This is an automated email from the ASF dual-hosted git repository.

altay pushed a commit to branch aaltay-patch-1
in repository https://gitbox.apache.org/repos/asf/beam.git

commit 9b043c983473eeb3a0b508c18426412dc67a0a58
Author: Ahmet Altay <aa...@gmail.com>
AuthorDate: Wed Nov 6 11:11:50 2019 -0800

    Increase overhaed budget for test_sampler_transition_overhead
    
    This test is flaky when it is run under load. Increasing overhead budget as a temporary fix.
    
    A better solution would be to change the test structure to not depend on specific timings.
---
 sdks/python/apache_beam/runners/worker/statesampler_test.py | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sdks/python/apache_beam/runners/worker/statesampler_test.py b/sdks/python/apache_beam/runners/worker/statesampler_test.py
index 176f6e5..b522294 100644
--- a/sdks/python/apache_beam/runners/worker/statesampler_test.py
+++ b/sdks/python/apache_beam/runners/worker/statesampler_test.py
@@ -117,6 +117,11 @@ class StateSamplerTest(unittest.TestCase):
     elapsed_time = time.time() - start_time
     state_transition_count = sampler.get_info().transition_count
     overhead_us = 1000000.0 * elapsed_time / state_transition_count
+    
+    # TODO: This test is flaky when it is run under load. A better solution
+    # would be to change the test structure to not depend on specific timings.
+    overhead_us = 2 * overhead_us
+    
     logging.info('Overhead per transition: %fus', overhead_us)
     # Conservative upper bound on overhead in microseconds (we expect this to
     # take 0.17us when compiled in opt mode or 0.48 us when compiled with in