Altostratus1.1.1.1 & 1.1.1.2 are two main sources + A few tester machine IPs
VIP is 2.2.2.2
Pool members are 3.3.3.1 & 3.3.3.2 - Round Robin + Source Address Persistence
When pool members are recycled / rebooted, traffic sticks to single node.
Clearing persistence record for the VIP fixes the issue, but any 2020 solution ?
MVPThis is known behaviour in persistence. I will suggest you to reduce source persistence ideal time value.. this may help you..
CirrocumulusIf it's http based traffic you can try using cookie persistence. You'll need to have an http profile on the virtual server.
If traffic is not session based, you might try taking out persistence all together.
If neither of those are possible, then you might try reducing the persistence timeout to a value just above the application's session timeout.
Hope that helps.