When queuing LLD rule value the LLD manger discards it if it matches the last value, since there is no need to perform 2 identical LLD discoveries in row. However since the rule comparison is done by simply comparing the text values - it will give false negative if the data contains also metric results. While less common case - it also can give false negative if the 'lld rows' are in different order.

This can be improved to comparing not all of data, but only the values defined by lld macros and the first level values with valid lld macro names. To do this
1) manager will send value to worker
2) if there are no more values of the same lld rule - manager will send 'process' command to worker
3) if there are more values:
3.1) manager will send next value to worker asking to check it for duplicate
3.2) worker will compare values using extracted macros and values
3.3) if values differ, worker will send to manager 'keep' response and start processing rule
3.4) if values are the same, worker will send to manager 'discard' response and wait for either the new value to check or 'process' command
4) manager will check the response received from worker:
4.1) 'discard' - manager will discard the last value and return to step (2)
4.2) 'keep' - no special action