Hi John,
The Veeam product they are referring to is a piece of backup software. In my opinion Veeam Backup and Replication is a great product but it absolutely can tank a VMs performance if it kicks off a job at the wrong time.
Back to the issue at hand. Reading through your posts my first suspicion is that the SQL server is underprovisioned. The standard GP requirement for RAM is far from the actual requirement when running MR on the same SQL server. My company went through crippling GP slowness while we got this sorted out. Admittedly this only happening at 4:30 on Fridays does make this suspect.
A few things you can try to narrow down the issue.
1. You can setup scheduled tasks to stop and restart the MR Services prior to the normal outage window. If the issue persists this is not MR related.
2. See if you can get more RAM allocated to the SQL VM. My current setup has 32GB allocated to it. We have a single company database that is around 40GB and SQL and MR share the same server.
3. Ask your backup Admin (DBA? some of us wear many hats) if there is a weekly backup job that starts around this time. A weekly backup is often a full backup which will last longer and be more resources intensive than a daily incremental.
4. Might as well remove the DM if you are only using the legacy setup.
Good luck!
Kirk