[BEAM-12603] Add retry on grpc data channel and remove retry from test.

[y1chi]

Please add a meaningful description for your change here

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Choose reviewer(s) and mention them in a comment (R: @username).
• Format the pull request title like [BEAM-XXX] Fixes bug in ApproximateQuantiles, where you replace BEAM-XXX with the appropriate JIRA issue, if applicable. This will automatically link the pull request to the issue.
• Update CHANGES.md with noteworthy changes.
• If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI.

[y1chi]

R: @TheNeuralBit

I was able to get 100 runs successfully with this patch, do you mind take a look and help me validate and decide if this is an acceptable fix?

[codecov]

Codecov Report

Merging #17537 (3405928) into master (0a01fbe) will increase coverage by 0.03%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master   #17537      +/-   ##
==========================================
+ Coverage   73.85%   73.89%   +0.03%     
==========================================
  Files         691      691              
  Lines       91255    91547     +292     
==========================================
+ Hits        67396    67645     +249     
- Misses      22626    22669      +43     
  Partials     1233     1233              

Flag	Coverage Δ
python	83.68% <100.00%> (+<0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...ks/python/apache_beam/runners/worker/data_plane.py	87.57% <100.00%> (-1.64%)	⬇️
...ks/python/apache_beam/runners/worker/sdk_worker.py	89.09% <100.00%> (-0.45%)	⬇️
.../python/apache_beam/testing/test_stream_service.py	88.09% <0.00%> (-4.77%)	⬇️
sdks/python/apache_beam/coders/row_coder.py	94.49% <0.00%> (-2.65%)	⬇️
sdks/python/apache_beam/runners/common.py	87.94% <0.00%> (-2.33%)	⬇️
...che_beam/runners/interactive/interactive_runner.py	89.43% <0.00%> (-1.41%)	⬇️
sdks/python/apache_beam/io/source_test_utils.py	88.01% <0.00%> (-1.39%)	⬇️
sdks/python/apache_beam/io/localfilesystem.py	90.97% <0.00%> (-0.76%)	⬇️
... and 9 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0a01fbe...3405928. Read the comment docs.

[TheNeuralBit]

Thanks! I think this workaround is preferable to retrying the entire test. A couple of questions/concerns:

• Are we sure it's safe to retry these UNAVAILABLE responses? The [gRPC docs note "it is not always safe to retry non-idempotent operations."
• I think this is a better workaround, but it's still a little concerning that we don't understand the root cause - any thoughts on how we could dig deeper?

[y1chi]

Thanks! I think this workaround is preferable to retrying the entire test. A couple of questions/concerns:

• Are we sure it's safe to retry these UNAVAILABLE responses? The [gRPC docs note "it is not always safe to retry non-idempotent operations."
• I think this is a better workaround, but it's still a little concerning that we don't understand the root cause - any thoughts on how we could dig deeper?

I believe UNAVAILABLE in this cases means that the underneath tcp connection was broken (not sure why and unclear how to debug that) before the request is handled so that means it should be retriable(had more than 250 runs and didn't see any side effect like getting wrong results after adding retry). I enabled the GRPC debug log but didn't find anything interesting also.