Well, I was aiming at a different aspect.
I was not precise enough; the VM itself had around 25 GB available in total for the mdf and ldf files.
Once we hit that, it stopped cold. The problem was it was a development database so there was a lot of churn going on, growing the ldf beyond what IT ever thought we'd need. Our process involved downloading huge files from an Oracle OLT database for analysis. One or two of those ETL processes filled up the log file rapidly.