Monitoring the performance
of a ClearCase installation can be time consuming, and it is not
straightforward to get significant values and their
ClearCase in itself depends strongly on the server
hardware, operating system and the network, and any of the pieces can
become that bottleneck, that decrease the user satisfaction.
Most of the ClearCase
performance studies shows that it quite often is the selected
work-flow, eventually I/O heavy triggers that are causing the pain,
unless the scripts are written specifically with performance in
Assuming you've taken care of trigger scripts, there
isn't much you can do, MVFS and View caches can sometimes be
adjusted, but we have never seen significant performance improvements
that made users happier than they were before.
But the main
thing to monitor is the network and a couple of key indicators on the
servers, which we will come back to a bit later, what will be really
interesting is to monitor these values over time and the ability to
configure alerting if any monitored value exceeds a given threshold.
With that in mind, we
created Monkit and CCAnalyzer.
is a Jenkins plugin for
presenting properly formatted xml data, please see Monkit's page.
client-server utilty that reports findings in the format Monkit expects.
CCAnalyzer is distributed as a java jar file, - you can install it as a Windows Service.
That can provide you graphs and history like these examples:
This is from a ClearCase multisite installation, we are monitoring 3 sites.
These charts shows the history of some disk related counters from one site, where are monitoring 2 servers.
Same site as above, but we are monitoring the some memory and processor counters on one more server.
Take a note of the horizontal red and blue lines, the are indicating warning levels for that specific counter. If a sample, meaning one measurement on one of the counters is above the threshold, a mail is sent to the administrator.