Thanks!
I will try and add retry functionality.
On Apr 6, 11:50 pm, Jeff S wrote:
> Hi Marcus,
>
> The error rate makes me think these are transient errors, so it should
> be safe to retry. Be sure to catch the specific exception, since there
> are other reasons that a datastore get might fail
Thanks!
I will try and add retry functionality.
On Apr 6, 11:50 pm, Jeff S wrote:
> Hi Marcus,
>
> The error rate makes me think these are transient errors, so it should
> be safe to retry. Be sure to catch the specific exception, since there
> are other reasons that a datastore get might fai
Hi Marcus,
The error rate makes me think these are transient errors, so it should
be safe to retry. Be sure to catch the specific exception, since there
are other reasons that a datastore get might fail more consistently.
(like an invalid key name ;-)
Timeouts can also indicate datastore content
Hello Jeff!
Thanks for your reply.
I would estimate that it is around 0.04% of the calls that fails.
Would it help to catch the timeout exception and do a retry directly
again or would that also fail?
I am mostly worried that the data from the users will go lost.
/Marcus
On Apr 4, 3:02 am
Hi Marcus,
You are correct that get_by_key_name does tend to be a very efficient
query, but it is possible to see a timeout though this is usually
quite rare as a percentage of overall requests. Do you have a feel for
what percentage of these get operations are timing out?
Thank you,
Jeff
On A