That means when the CLR started, the AppMon agent wasn't able to instrument all of the classes and packages. Sometimes this just happens - it can be a fluke on the network during startup time. If you go to that agent in your agents overview, check the "Average Instrumentor Response Time" and max time. We want these numbers to be as low as possible as to not hold up the CLR from starting because of AM instrumentation. If they're quite high, over a second, the agent distance from the collector could be an issue. If they're low, <100ms, we may have to dig in further. We recommend placing the collector within 2ms ping of the agents to keep this number low.
If you can, I'd first try restarting those .NET processes and see if they work the second time.
+1 for what Brett said. One thing to add: High ping time between the agent/Collector will certainly result in instrumentation slowdowns, however throughput is also a factor. So just because ping time is low does not mean the link is acceptable. Therefore the "Avg Instrumenter Response Time" is probably the most credible metric to determine if the Agent/Collector link is causing such a problem.
Also one thing to add here. I just got the same error message for a Web Server Agent. The real root cause however was that the agent name didnt match any system profile. Normally that would lead into a message saying "Couldnt find matching system profile" - but as I said - I just had this issue today on an installation where I saw your error message even though the problem was incorrect name matching!