Hello! Every so often, my async_bulk load fails with a Connection Timeout. I have my timeout parameter set to 60; I had set it arbitrarily high, but it didn't pass code review. I can't just wrap the call in tenacity as I can't really expect it to rewind the generator.
What's the best practice for selecting a timeout?
I don't know how often it fails. Each of my load jobs does a few hundred 500-document batches, and one in eleven of the jobs fail, so call it one in every five thousand times.
Also... the docs say async_bulk will only retry on 429 (too many requests) response status. Is that the case? Will it retry on Connection Timeout?