How to avoid the ‘too many measurables’ event message

The ‘too many measurables’ event messages are logged when a monitoring data producer has more than 500 measurables. When the event is logged, one can check the list of measurables created by the producer mentioned in the message, for example if ‘Too many measurables for: ProcNetDev’ event message is logged, then the list of measurables created by ProcNetDev can be checked by running this cmsh command:

% monitoring setup get procnetdev measurables

If the list of measurables doesn’t have any unnecessary measurables, then it’s possible to increase the limit for maximum number of measurables per producer by setting the cmdaemon advanced config parameter MaxMeasurablesPerProducer to a suitable value. The suitable value for the parameter can be determined from the number of measurables.

% monitoring setup get procnetdev measurables | wc -l

For example, if the current number of measurables created by the producer is 550, the value of MaxMeasurablesPerProducer can be set to 600. If the /cm/local/apps/cmd/etc/cmd.conf file on the head node has these advanced config parameters set:

AdvancedConfig = {
                   "VirtualCluster=1",
                   "ExtraCookieArguments=SameSite=None",
                 }

Then “MaxMeasurablesPerProducer=600” can be added as shown below:

AdvancedConfig = {
                   "VirtualCluster=1",
                   "ExtraCookieArguments=SameSite=None",
                   "MaxMeasurablesPerProducer=600",
                 }

After adding the advanced config parameter, the cmd service on the head node has to be restarted. 
If there are 2 head nodes (because of HA), then the configuration change and cmd service restart have to be done on both of the head nodes.