Hi.
I'm also having the same issue. Not only with Azure, also with Okta SAML.
I've enabled the TRACE logs for pac4j, as the error suggests the issue instant is indeed old, I've seen instances issued 2-3 days before, one time even a month old.
So the error is correct. What I don't understand is how this should be resolved. Can I instruct the IDP in the initial request not to reply with old instances? if not, can I redirect back to the IDP after the failure to refresh the auto instant?
or should I just increase the value to a value large enough?
The strange thing is when checking the azure session timeout it is 24 hours, yet I see instances issued earlier.
Thanks