During a long drush-running job (that sends emails via SES) the temporary credentials expire. The amazon_ses module catches that and re-creates a new SesV2 client, which in theory obtains new temporary credentials. In practice it seems to re-use the credentials in the key/value store, even though they *should* have expired at that point.
That is because the keyvalue.expire get() call sues REQUEST_TIME, which never changes during even a multi-hur drush run.
Send out a *lot* of notification emails, see errors after an hour :-)
Add a method to allow the temp credentials to be dropped form the key/value store, so that they get refreshed even if REQUEST_TIME is <<<< now().
Draw the rest of the owl.
A new function in the ProfileInterface.
Active
2.0
Code
Oh, except I have no way to call that new method from an AwsClientFactory, I think. So it's not actually useful.