I have been doing a lot of research on Veeam best practices and wanted to get some other user's feedback. I have a multi-tenant cloud (both Hyper-V and VMware) that will grow fairly large over the next year with upwards of 10 hosts and possibly thousands of VMs connected to on-site SANs (with off-site redundancy for backup jobs). I'm looking into options for providing custom retention policies to for different tenant needs in the environment and was hoping to find some wisdom here. I've read in several other posts and in some Veeam documentation that is makes sense to group "like" VMs into replication/backup jobs for compression and deduplication considerations, and I've also heard others mention to keep the job count low for overhead and management considerations.
My question is, how many jobs is too many? Will configuring a new custom job for each tenant's environment (<10 VMs) cause serious IOPS and network performance issues (all will be Forward Incremental with 1 weekly synthetic)? Would it make sense to group tenants based on how many snapshots they want to retain (say, have 3 backup jobs, one for minimal, medium and large retention policies; something like 3, 14, or 30 daily snapshots)? Or should I look at limiting individual tenant retention policies and standardizing the number of restore points to allow, and instead group jobs by guest OS/VM purpose?
Any feedback is helpful. I appreciate your time in offering a thoughtful response.