capthiron commented on PR #24364: URL: https://github.com/apache/beam/pull/24364#issuecomment-1330785836
Unfortunately we are currently facing scalability issues with either this or the previous version with the groupBy step. Do you have any Idea why this approach would not scale? We recently ran a comparison with the python-based Bigtable connector which in general was slower but then took advantage of dataflow's autoscaling. This scaled up to 28 workers and the throughput pretty linearly which lead to a reduced the duration of an import with over 110 million rows by a factor of 10. Do you have any ideas where the problem could be? -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
