DX Unified Infrastructure Management

Still working on this POC :-)

There is a requirement to have different thresholds set for various metrics at different times of the day. I was hoping that Dynamic Thresholds would be able to satisfy this requirement but customer requires more flexibility than this. They effectively want to have a schedule for each metric to cover:

1. Set of thresholds per metric for specific time periods e.g. metric < 1000 during peak hours and metric < 500 during off-peak hours.
2. Set of thresholds to cater for expected and predictable peaks e.g. marketing periods - metric < 1500.

Above is just an example and not exhaustive.

By the way, the SQLSERVER probe seems to cater for the above - it has a schedule that can be applied to each metric - but that's not typical. I was thinking that the above could be done via custom code but no idea how long to quote for a services engagement.

Has anyone come across this requirement and developed a solution?

Alternatively any good arguments why you wouldn't want to do above?

Thanks

Ray

I had a similar requirement awhile back and Jim was nice enough to throw together this solution - it shouldn't take much effort to modify it for your own environment:

**********************************************************************************************************

Here is a script that will do what you asked. You will need to replace the NimBUS address with the correct one. You will also need to copies, one for each desired interval change, then run the scripts with the scheduler.

Remember to make sure the NAS Probe access is set to admin:

-- Script to change the polling interval of a probe
-- Replace the NimBUS address here with the correct one for the customer
addr = "/JMC_Development/jchristen-lt2/jchristen-lt2/controller"

cmd = "probe_config_set"

-- the required arguments to a nimbus.request() are required to be sent in a PDS object
-- the probe_config_set command needs the probe name, config file section, the key, and the value
myArgs=pds.create()
-- probe name
pds.putString(myArgs,"name", "interface_traffic")
-- config file section
pds.putString(myArgs,"section", "setup")
-- key
pds.putString(myArgs,"key", "interval")
-- value
pds.putString(myArgs,"value", "30")


nimbus.request(addr,cmd,myArgs)

pds.delete (myArgs)

-- Deactivate the probe

cmd = "probe_deactivate"

myArgs=pds.create()
-- probe name
pds.putString(myArgs,"name", "interface_traffic")

nimbus.request(addr,cmd,myArgs)

--Active the probe
cmd = "probe_activate"
nimbus.request(addr,cmd,myArgs)

pds.delete(myArgs)

Hello Ray,

This is very common requirement in prodcution envrionment where the system load varies during business and non-business hours, especially if it's a customer facing server or a batch processing server.

You can implement the solution using either:

a) Scehduler and central configuration repository

or

b) Multiple config sets (locally on monitored server) and automated action to swap the config file.

Thanks,

Gyan