Has anyone else ever had instances of workflow actions just not firing when they're supposed to? It happens super infrequently, but I have two instances documented with support over the past ~14 months and another one just now (I think).

I’ve definitely had this before

Yeah, completely unable to replicate the behavior in prod, sandboxes, or NS support's cloned environments, but the examples of it not working are clear as day in the logs where some actions are firing and others are not, all with the same triggers. I have a defect for this, 417298, but it's currently assigned U6 and like I said, has been sitting with no action for over a year now, and support seems to treat it like "well, we can't replicate it, so there's no problem". I mean other companies also use workflows to do approval flows and other internal controls that are relied upon for sox/audit, right? I get that it's really hard to fix something you can't easily replicate, but I feel like at this point this should be a hair-on-fire, you-should-be-failing-your-SOC-certifications problem if parts of workflows can just silently fail for no reason.

@James Morris this is reminding me of a problematic SFTP server connection we had which we relied upon once per day. It would intermittently fail. We deployed a separate script just to ping the SFTP server every 15 minutes and log the response to a text file. Eventually we had more data and could demonstrate the issues. issue was the vendor’s SFTP server itself not netsuite. So I wonder if you can create something that can load test this issue to your sandbox and test it 12 times a minute for the next 24 hours to generate enough replications .

That's a really good suggestion. I'm a little worried I won't be able to properly capture what's actually causing the issue in my test since it happens so infrequently, but worth a shot for sure.