Another valuable tool for checking links is Link Quality at a Glance from Hans van der Graaf.
Edit $XYMONHOME/etc/jmxstat.ini, as a minimum you will need to modify the following in the default section :- NCV_Threads="Live Threads: GAUGE, Peak: GAUGE, Daemon Threads: GAUGE, Total: DERIVE" NCV_JVMHeap="Heap Size: GAUGE, Committed: GAUGE, Max Heap Size: GAUGE" NCV_Classes="Classes: GAUGE" NCV_Web App="active Sessions: GAUGE,session Counter: GAUGE,max Active: GAUGE,session Max Alive Time: GAUGE,session Average Alive: GAUGE,rejected Sessions: GAUGE,expired Sessions: GAUGE" NCV_Thread Pool="current Thread Count: GAUGE,current Threads Busy: GAUGE" NCV_Service="bytes Received: DERIVE,bytes Sent: DERIVE,error Count: DERIVE,request Count: DERIVE"[Threads] # java.lang:type=Threading TITLE Java Resources YAXIS Threads DEF: Lve=Threads.rrd: Live Threads: AVERAGE DEF: Pek=Threads.rrd: Peak: AVERAGE DEF:r=Threads.rrd: Total: AVERAGE CDEF: Tot=r,60,* LINE2: Lve#00CCCC: Live Threads LINE2: Tot#FF0000: New Threads COMMENT:\n GPRINT: Lve: LAST: Current Threads \: %5.1lf%s (cur) GPRINT: Lve: MAX: \: %5.1lf%s (max) GPRINT: Lve: MIN: \: %5.1lf%s (min) GPRINT: Lve: AVERAGE: \: %5.1lf%s (avg)\n GPRINT: Tot: LAST: New Threads/min \: %5.1lf%s (cur) GPRINT: Tot: MAX: \: %5.1lf%s (max)\n [Classes] #java.lang:type=Class Loading TITLE Java Resources YAXIS Classes DEF:count=Classes.rrd: Classes: AVERAGE LINE2:count#00CCCC: Classes Loaded GPRINT:count: LAST: %6.1lf (cur) \: GPRINT:count: MAX: %6.1lf (max) \: GPRINT:count: MIN: %6.1lf (min) \: GPRINT:count: AVERAGE: %6.1lf (avg)\n [Service] # Catalina:type=Global Request Processor,name=http-8080 TITLE Java Resources YAXIS Requests/sec DEF:in=Service.rrd:bytes Received: AVERAGE DEF:out=Service.rrd:bytes Sent: AVERAGE DEF:r=Service.rrd:request Count: AVERAGE CDEF:req=r,1,* -X0 LINE2:req#FF0000: Requests COMMENT:\n GPRINT:req: LAST: Requests/s \: %5.1lf%s (cur) GPRINT:req: MAX: \: %5.1lf%s (max) GPRINT:req: AVERAGE: \: %5.1lf%s (avg)\n GPRINT:in: LAST: Received bytes/s\: %5.1lf%s (cur) GPRINT:in: MAX: \: %5.1lf%s (max) GPRINT:in: AVERAGE: \: %5.1lf%s (avg)\n GPRINT:out: LAST: Sent bytes/s\: %5.1lf%s (cur) GPRINT:out: MAX: \: %5.1lf%s (max) GPRINT:out: AVERAGE: \: %5.1lf%s (avg)\n [JVMHeap] TITLE JVM Heap Memory YAXIS Memory DEF:kused=JVMHeap.rrd: Heap Size: AVERAGE DEF:ktotal=JVMHeap.rrd: Committed: AVERAGE CDEF:bused=kused,1024,* CDEF:btotal=ktotal,1024,* CDEF:free=btotal,bused,- CDEF:used=bused -l 0 -b 1024 AREA:used#0000FF: Used Memory AREA:free#00FF00: Free Memory: STACK COMMENT:\n GPRINT:used: LAST: Used \: %8.1lf%s (cur) GPRINT:used: MAX: \: %8.1lf%s (max) GPRINT:used: MIN: \: %8.1lf%s (min) GPRINT:used: AVERAGE: \: %8.1lf%s (avg)\n GPRINT:free: LAST: Free \: %8.1lf%s (cur) GPRINT:free: MAX: \: %8.1lf%s (max) GPRINT:free: MIN: \: %8.1lf%s (min) GPRINT:free: AVERAGE: \: %8.1lf%s (avg)\n [GCInfo] TITLE Garbage Collection Activity YAXIS # collections FNPATTERN GCInfo.(.
It is modelled on the monitor for Weblogic ( but jmxstat should work for any JVM that exposes the JMX interface, especially Tomcat, but also JBoss and Weblogic.
The jmxstat monitor provides the same reporting and alerting facilities as beastat and is configured in a very similar way so anyone familiar with beastat should be comfortable with the installation. Copy the 3 files jmxsh-R5.jar, and to the Xymon extensions directory $XYMONHOME/ext 2.
You can subscribe to the list by sending an e-mail to hobbit-announce-subscribe@ Deadcat is mostly Big Brother scripts, but most of them will work on Hobbit without too much fuss. hobbitping normally stops pinging a host after receiving a single response, and uses that to determine the round-trip time. Or perhaps more frequently (a critical connection once per minute or less). It defines how long a status message is considered valid (i.e. Search "LIFETIME" on the bb man page for a bit more info. The defaults are 5 minute testing intervals and 30 minute status lifetimes, so 5 misses are accepted by default.
The Shire Project (Xymonton) https://wiki.xymonton.org/Shire Project is specific to Xymon(Better than Deadcat) A. The lifetime should be adjusted for the acceptable number of misses, otherwise false positives might occur.