So a helix 10 and a LVS32/34/blackbox/10" headunit will combine for about 5-6 amps. A group 31 AGM is about 100 AH. Since you have two in parallel, you've got about 200 AH capacity which should be WAY more than you'd ever need. That said, 12.2V is around 65% full. I'm thinking there are a couple things in play here.
1- the connections sound like they are all on one battery with the second battery in parallel off the back, as opposed to both in parallel but with a primary connection on each battery. What is happening is that you're working the battery that the main connections are on a lot harder than the second one backing it up. Long term, I'd swap one of the terminals if you can so that they are drawing through both batteries the same, if for no other reason than evening out the wear on the two. If the Humminbird is connected to the 'primary' battery in your setup and the Garmin to the backup (for lack of a better term), that could explain why the HBird fluctuates so much and the garmin doesn't. To that end, for the next trip out I would separate the batteries and run one as a dedicated electronics battery to see if that solves your issue. It should be a simple swap and easy to test.
2- It could also be that the units have different sensors and processing of the voltage sensor readings. So the same battery might show differently on each unit. Similarly, the voltage at the terminals won't be the same as at the unit due to losses in the wires between the two. A long skinny wire (humminbird helix power cables for instance) will drop move voltage than a short fat wire. If you have both on the same battery and no motor/alternator playing with voltage, then you can get a gauge on how they are vs each other.
3- Maybe most importantly, you're getting a low voltage alarm but has the unit every shut down or given problems? The min voltage spec is 10.8V. Do you just have your low voltage alarm set at a high number that doesn't make any real difference in use?