You can redesign existing solution using workflow action under "Flow Control " called Condition Parallel. Each branch should be set up against the flag which would further be validated down in the flow thus will help you understand the pattern when the service is getting timeout. Please ensure you use Delay action in each branch of at least 10 seconds starting from second branch. . Each branch should be broken based email number so if 1000 email are sent then set up at least 10 branches to start with. Then it can be increased or decreased so forth based on the understanding that we develop.

I am working with the task error handler and a try/catch loop.

How do I specify a "task error handler" in the catch loop, there are a bunch of error types to pick from however I cannot locate the specific task error handler I create in workflow admin.

Perhaps I am missing something?  The docs describe this as an option but provide no example configs.

Hi, I consider that this is not a solution, it is only a patch (badly done), even this How do I do for LF tries to connect to Server SMTP again? because once does not send ONE email it never send again until restart service 

Thanks.

In my try/catch loop I am seeing the catch loop triggering and I can correlate that to a timeout connecting to the SMTP service (mandrill).  These are rare errors but enough that I am seeing them.

On the catch we are re-sending the email/attachment so that effectively 'tries" again and thus sends the email that timed out.

So, this approach does seem to work as far as resending timed-out SMTP sends.