There’s a known issue with the Azure AD DS Health Monitoring Agent, which is a part of the Azure AD Connect Health offering from Microsoft.
I’m a big fan of this service, which after installing a small agent on each DC, will alert you of any issues such as replication failing, or a DC unavailable.
However, there’s a problem with how the agent handles its temporary files. As covered on this TechCommunity post, the utility creates a lot of temp files in C:\Windows\Temp locally on each DC. They’re 1/2KB each, but I see around 288 daily being generated. These are never cleaned up.
One one domain controller, since I’ve been running the utility from the 16th September 2016, there are now ~133,000 temporary files. The actual size of these log files is a small 90mb, but the space on disk due to how allocating blocks works, takes up 519MB. I’m going to assume there’s many factors that can change the size and number of log files.
Many people will have small drives for their DCs, and also having lots of files in a folder can cause weird performance issues.
The files are in a format such as 20160915T024226Z-20160915T031125Z-SERVERNAME-6acbd4cb99a1448d848298a59b6fc6e2.json.gz – so it’s easy to set up a daily scheduled task to delete anything older than a day. There’s a couple of examples on how to do this here.
Microsoft has advised this won’t be fixed anytime soon (at least Q3 2018 is what I’ve heard), so it’s worth checking out that C:\Windows\Temp folder and even doing a one time delete if it’s full of log files!