These parameters determine how the cluster and instances manages the memory used to store the data for the historical monitoring tables:
deadlock pipe max messages
errorlog pipe max messages
sql text pipe max messages
plan text pipe max messages
statement pipe max messages
You can configure these parameters globally for the cluster and individually for each instance. These parameters allocate memory for the pipe. An instance can dynamically add memory to the pipe but cannot dynamically remove memory from the pipe, so if you reduce the size of the parameter, you must restart the instance for the new pipe size to take effect.
Below are some algorithms for determining the size for the parameters.
For an individual instance, the memory required for the each pipe configuration is:
configuration_value X number_of_engines
To globally set the memory for each pipe configuration:
configuration_value X number_of_engines X number_of_instances
If you have set the value for pipe configurations differently for each instance, then the amount of memory required for the cluster is:
(instance_1_value X number_of_engines) + (instance_2_value X number_of_engines) +. . . + (instance_n_value X number_of_engines))