How I can try to use v3?
melkayam92 opened this issue · 4 comments
Currently, I am using the bigtable_v2 async client, which generally works well. However, I have observed a spike in latency every hour. I suspect this may be due to the GRPC channel closing after 1 hour. Upon reviewing the v3 code, I discovered the presence of a grpc connection pool, which seems promising for resolving my issue.
I attempted to install the branch using pip, but it was unsuccessful. Do you have any suggestions on how I can test it?
We're planning on cutting the first alpha release at the end of the week, so I'd recommend waiting for that. I believe it will be released as a developer pre-release on pypi, but you should also be able to install builds directly from the repo by cloning it and installing with pip install -e . If you're having issues with that let, me know (currently you also need to clone and install the python-api-core sub-module, but that won't be required after the alpha this week)
If you decide to try out the alpha and have any feedback, please reach out!
As for the latency spike you're seeing, it could be due to the channels needing refresh, or it could also be due to oauth credentials expiring and needing refresh, which I believe also happens ~hourly. The channel pool in the alpha should address the first one, but auth token refresh is coming later.
Thanks @daniel-sanche , I will wait for the alpha.
For now, I have implemented a workaround where I warm up a different client every 50 minutes. It seems to have resolved the issue with the spikes.