Hey Egor,
According to the guidance it sounds like you are maxed out. A two-node partitioned DDB can handle 1PB backend data. That being said, the guidance is conservative - and many are able to vastly exceed the guidance depending on their hardware, configuration and profile of the data they are protecting.
There is no ‘hard limit’ - you can keep going until you see performance degradation.
Performance will slow down once the Query and Insert time on the DDB slows - the software will alert you when that happens or when you are approaching that limit.
Besides that, the obvious signs are approaching memory or CPU capacity, and If your jobs begin to exceed their SLA / backup window then that is a typical sign that you are reaching the limits and should consider adding an additional node.
@Damian Andre Thank you for the answer, at the moment we dont have any issue with DDB the I/O is at 5-14%, we do have some issue with iowait of the mediaagents. And the AUX copy getting slower en DDB verification two.