My team noticed that the defaults for azure storage client (particularly min/max_retry_delay) at 10ms/500ms are much lower than that suggested by Azure (0.8 seconds - 1 minute) respectively. Is there a reason there is such a discrepancy here? We are having throttling errors in our storage accounts which is causing a lot of query failures, and Azure suggests to up these values significantly.
Is there a danger to increasing these values by a notable amount?
I don’t use Azure personally, but I think it should be safe to change those values, but you do want to keep the still relatively low.
I suspect those values are kept low because you do also need to consider the default query timeout. If your query is timing out while the reader is waiting to retry it wouldn’t do you any good necessarily.
Also, if you have a ton of logs you might also consider using multiple storage, although I don’t know if it actually helps with Azure, and it’s certainly a big change that you’ll want to plan and test for.