GPT processing error subset area outside image - how to automate/check

Short version: Is there a way to check whether a subset area intersects a given Sentinel-1 image in an automated way (as part of GPT), so processing can continue for all subset areas?

I am using GPT (from the command line) to run a processing chain on a Sentinel-1 image, on several small tiles (we’re only interested in a narrow band along several rivers). In a first step we download all Sentinel-1 images which intersect the complete extent of all small tiles. The next step uses GPT to run several Sentinel-1 specific processing steps, followed by subsetting to one of several river tiles. This continues for all small tiles along the river. We have to tile the images, because we are using them in another software package which cannot handle entire Sentinel-1 images.

When we run the processing and the tile falls outside the raw Sentinel-1 image, GPT returns two messages. The first makes sense and is not the problem: WARNING: org.esa.snap.core.gpf.common.SubsetOp: No intersection with source product boundary.

The problem is that it then and returns something like: ERROR: Operator ‘SubsetOp’: Value for ‘Reference Band’ is invalid: ‘Gamma0_VV’". This causes the processing chain to stop and not continue to process the other tiles. I tried reordering the processing steps in GPT (putting the subset last), but another similar error always pops up.

Is there a way to dismiss or turn off the second validation error, which is simply a result of the subset area bring outside the limits of the Sentinel-1 image?

What I would like is for GPT to check whether a tile overlaps with a raw Sentinel-1 image. If it does, then do the full processing line. If not, skip the tile (or make an empty file). One way to do this would be for GPT to simply run the processing chain but without throwing a error (e.g. give a warning about subset area being outside image extent). Is this possible? If not, is there a better/other way to do this in an automated way?

In the past we have done the subsetting step using snappy, but we are trying to avoid using snappy since we had major issues every time we had to install it on another computer. We would really prefer to use GPT and some .bat files, if at all possible. Any ideas/suggestions are appreciated!