Installation and Operation
A tarball called agent-linux.tar.gz is available for download through the web interface. JRE 1.6+ is required and 2 daemons are installed and launched – a watchdog (presumably to ensure the agent daemon is running), as well as the agent itself. The default install directory is /usr/local/logicmonitor/agent. The logs directory can be found here as well. This shows the communication events between the local agent and the remote server. Communication is only outbound so you don’t have to worry about opening up ports in your firewall.
It’s expected that one agent will monitor several servers (for example one per data center). A nice add-on is the ability to set backup agents in case the main fails for any reason (leaving you suddenly without any monitoring at all). I have 2 agents for our production/staging environment (they backup each other), and separate agents for the build, backup and puppetmaster servers.
For us folks over here in Europe, LogicMonitor has a data center in Boston. I haven’t had any problems in lag or performance with either the agent or the web interface.
Writing your own monitoring scripts is this easy:
#!/bin/bash SERVICE=$1 if pgrep -fl $SERVICE then exit 0 else exit 1 fi
You can also easily clone and extend existing checks (definitely recommended over just going in and editing the default checks) for more customizations.
A very good question I got was what about existing Nagios datasets/checks (i.e. performance data). Reading through the documentation and website I really see nothing about this. I can’t imagine accessing and monitoring these files would be a problem for LogicMonitor, but some official feedback would be appreciated on this point!
Scaling and Pricing
Straight from the horse’s mouth we’re told it’s $300 for up to 25 servers (LogicMonitor pricing. Do the math and you get a very comfortable 40 cents a day / server.