Support timeout / retry on Table.read_row()

[mwarkentin]

Is your feature request related to a problem? Please describe.

We recently experienced an issue where a small number of requests to BigTable using Table.read_row() hung until we hit the default 60s timeout. This caused issues with our application including the kafka consumer being kicked out of the consumer group after not responding for 30s.

We want to set a more aggressive timeout for this function call (eg. less than 30 seconds) however we haven't been able to figure out a way - read_row() does not seem to support either timeout or retry args which are the documented ways to set a timeout on a request.

Describe the solution you'd like

Support table.read_row("keyid", timeout=5)

Alternatively, support setting a new global default timeout via environment variable or config file.

Describe alternatives you've considered

Perhaps we could refactor our code to use read_rows() instead and fetch a single row, but this feels hacky.

Additional context

• Existing support issue: https://console.cloud.google.com/support/cases/detail/v2/49879020
• Google account team members: Jason Shin, Scott La Rue

[mwarkentin]

@daniel-sanche @ayubun looks like this was tackled in #982 (THANKS!!)

Should I close this issue now? (I haven't had a chance to validate the update yet though)

[mwarkentin]

Ah, I see that release is not actually available yet. Will wait to confirm. :)

[daniel-sanche]

I just cut the release for v2.27.0, so I think we should be able to close this. Thanks!