The DPM management pack is one of the better management packs put out by Microsoft. It relies on the DPM eventlog to detect most of the issues should they arise.
However, when a lot of protected resources fail to backup over an extended amount of time, this eventlog tends to grow quite fast.
At this point it is possible that the event-generation of DPM ‘outruns’ the detection-speed of SCOM. When this occurs, this alert is raised:
This essentially means that the monitoring-agent on the DPM-server cannot keep up the pace.
A nasty side-effect of this issue is that the agent is ‘time travelling’ in the past, essentially catching and alerting on events that occurred hours, even days before. So in order to prevent this unwanted noise, it is important to give the agent some breathing room.
The easiest solution is to verify that no issue is currently ongoing on the DPM-system, and then clear the DPM-eventlog (after backing it up if necessary). This will ‘warp’ the agent back to the present time and allow event-processing to run normally. You can then close all false alerts, and the backlog-alert.