Is it possible to list all currently running agent machines and/or agent machines that do not have a currently running agent. Also would it then be possible to get notified of agents that go off-line.
To keep track of all agents that are currently running (and not running), the DSM Domain and Scalability Servers would have to poll the agent machines regularly to check if the agent is still running or alternatively the agent would have register itself regularly.
The current default is that the DSM agent registers itself once a day with the Scalability Server. If all agents were configured to register themselves every hour or every half hour then this would create a huge overhead for the network and the servers. For DSM agents on end-user PC's it is recommended to keep the default registration set at once a day.
If critical servers need to be monitored continuously then a Unicenter NSM set up should be considered with a NSM agent that can be configured such that any problem on the agent server is passed on almost instantaneously to the NSM management server.
To monitor the health of a large number of DSM agents or an individual agent consider the following options:
- Run a query that reports on all computers where the 'Last Agent Run' date is over a certain number of days. In DSM Explorer, right click 'Queries' and select 'New?'. Select 'Computers' and in Query Designer click on 'General Information -> Computer' and select 'Last Agent Run'. In the drop down field select 'Days older' and enter '7' to select all agent computers that have not reported in for over 7 days. Click OK twice and enter a name for the query. Then run the query and if there are any agents listed, investigate why they have not reported in for over a week.
- In DSM explorer an individual computer has the option of running 'Instant Diagnostics' (by rightclicking on this agent). The result will show if the DSM agent is up and running (and any of it's plugins) on this machine.