Well - to solve your issue with the very large Cluster - nworks 5.7 has the answer
In nworks 5.7 we are introducing the feature of 'Cluster Splitting' - this means that a single cluster job can now be automatically split into individual host jobs, and you can load-balance those jobs across two (or more) Collectors.
At the moment, although you are overloaded - 102% is not very serious overloading. Are you seeing any performance issues from the Ops Mgr agent on that Collector server? There is some tuning you can do (collection intervals, metric classes and multipliers and so on) that will improve the performance. Please see our Operations Guide for details, let me know if you have any questions!
Why you are seeing two different values (51% and 102%) for the same Collector I am not sure - could you open a case with our support for that? It's support at veeam.com
. From what you say on object count, 102% is the correct value.
Meantime - I think you can run at 102% load until we can release 5.7 - it is coming very soon (later this quarter; and I hope not too much later
Right now we have a beta of 5.7 - would you like to join the beta program? As with all betas, we don't recommend it for production use (possible upgrade issues to RTM, etc) - but we would love to get your feedback! And it could be your POC of 5.7, then you will be ready to deploy as soon as we release....
If you would like to join the beta, please post here and let me know!