Transaction Group large delays when Device Connection in ReconnectWait state

I am running Ignition 8.1.2 with 8 devices (AB PLCs) and 4 DBs (MSSQL) connected. There are 8 projects running with each project consisting of a Transaction Group setup to run multiple Stored Procedures on the DB for each PLC. Each transaction group may overlap on DBs.

Each project references the same default Tag provider, and within this, there's at least some tags for every PLC -- even projects that don't reference that specific PLC for that Transaction Group.

Under normal condition, things are fine, transactions all handled correctly in timely manner.

However, today witnessed when one of the PLCs (PLC A) drops out, another PLC (PLC B) project/Transaction Groups basically come to a halt, and transactions wouldn't complete. Once PLC A comes back online, PLC B Transactions move at the ms speeds again. This was repeatable.

I've only witnessed this PLC A / PLC B interaction so I'm not sure if any other Transaction Group is halted for the other PLCs when a device is lost elsewhere.

When I say PLC A drops out, I mean goes into the ReconnectWait state. The timeout on PLC A is set to 2000:

I'm looking for help on what to troubleshoot if this shouldn't be happening, or if this is correct per the current design and how to improve if it's a design flaw.

Lots of bugs have been fixed between 8.1.2 and the current 8.1.25. You should at least test a newer version.