I figured out how to adjust it. I raised it to 90% for now. We have ordered new storage nodes, but I need to make the array work until they arrive and are added.
If the GSAN is approaching capacity and the disk partitions are rising above 85% and preventing GC from running then the cause will typically be either
a) Due to a large amount of changed data since the last validated checkpoint
b) A full HFScheck failed or did not take place for more than a day
Either scenario will cause the checkpoint overhead to increase and on a full-ish system where the disk partitions are already around 65% full you've a maximum of 20% of the data partition space for that overhead
Raising the disknogc limit to 87% is acceptable on full systems but if it needs to go any higher than that you may be treating the symptoms whilst the root cause remains at large.
Couple of items..
1) Make sure you get a full HFScheck *every* day
2) If your Avamar version uses HFScheck schedules, change them to run full HFSchecks ever day
3) Use the capacity script to figure out if you have any particularly high change rate or check for brand new clients that might be causing a large amount of change on the system
4) If in doubt, open a support case