When performing an automated snapshot restore in Ops Manager, the MongoDB Agents shut down all MongoDB cluster members while the snapshot(s) are downloaded to the dbPath's.
During this phase of the plan, the MongoDB Agents continue to attempt network connections to the cluster members; as frequently as ~4 per second per mongod. This traffic may pose problems for certain cluster deployment designs. Of note, clusters that are "micro-sharded" (multiple cluster members residing on the same physical hosts) and clusters with members in a remote datacenter can receive an amplification and/or excessive amount of connections relative to what the network can accommodate.
This feature request is to make the MongoDB Agents state aware. If the agent plans call for shutting down the cluster members, skip or avoid frequent network connectivity tests until the download of the snapshot(s) completes.
This enhancement will also have the benefit of reducing spurious error logging during cluster maintenance and restore from backup.
