One of the more important things of Tomcat Monitoring is to keep an eye on the memory usage. The solution for most Tomcat/Java Server Problems is...

More Memory!

Some say there's only one better thing than more memory for a Tomcat Server:

Even more Memory!

And for the monitoring part: I suggest you to have a more broader look at your problem. When you have a certain message "...bla...broken pipe...", then you may know it already "Ah this one again" or you dig into your system until you figure out the root cause of the problem. What helps a lot in discovering the reasons are system performance data.
If you have those data, you'll discover from a look at those graphs what is normal for your system at what is not and above all: you'll discover all sort of changes before a critical state is reached. You'll learn what the impact of certain technical changes is, when you see the data coming in after that.

I suggest you to have a look at some monitoring system. As I assume you work in some type of company, there's probably already a monitoring system installed. I would expect most monitoring systems to be flexible enough to add your specific own checks there or use the features already available there. If no monitoring system is available to use, I recommend monit as an easy one ore check_mk as a more powerful approach - the latter needs to be on a separate best dedicated system(vps is enough). (Maybe there's a better solution for solaris available, but I do not know that).

There are some checks available for Tomcat/Java and check_mk. The client(=your tomcat server) part consists of a bash script most of the times. Sometimes, so chances are good that they just work at your server. Look here: Catalog of check plugins

And of course that one you're after(log file pattern matching) may be an import issue too. Most monitoring applications do have such features available.