Companies around the globe rely on HandySoft’s BizFlow® solution to meet their business process needs. However, in complex deployments, the BizFlow environment itself can pose significant monitoring and management challenges for the IT professionals responsible for its care and feeding.
Over the years, Absolute Performance has worked with a number of clients to optimize their business-critical BizFlow environments. During these engagements, we’ve noticed three common early indicators of problems that can lead to poor performance or outages in complex BizFlow deployments:
- Log Analysis
- Blocking Processes
- Queue Analysis
Keeping tabs on these three "hot buttons" is critical to the health of your BizFlow application and its underlying infrastructure. In this article, we’ll focus on analyzing and cleaning BizFlow error logs.
The Importance of Vigilant BizFlow Log Analysis
BizFlow log analysis is important for two key reasons:
- The BizFlow application generates a massive amount of log entries, and ...
- Many BizFlow performance problems initially present themselves in the log files
Often, people managing BizFlow environments gain a false sense of security by looking at the logs once and, after that, assuming everything is okay. This approach can quickly backfire because:
- Some common BizFlow errors only present themselves under heavy loads – say, at the end of the day or the end of the month
- In some cases, it’s the frequency of the errors over time (not the errors themselves) that matters
While the BizFlow application may behave normally during these time periods, keeping a close watch on the error logs could shed light on a potential problem before it affects the end-user.
Good Housekeeping for BizFlow Log Files
Regular cleaning of the log files is another important BizFlow management task. If one doesn't address errors and periodically purge the log entries, the log files can quickly grow to an enormous size. For instance, the ERA logs in BizFlow 9 can generate 5+MB per minute of data!
When the log files grow unchecked, the activity of logging itself can become a huge drain on the performance of your BizFlow environment. In fact, many IT professionals managing BizFlow environments aren’t even aware that some of these logs exist until the logs grow large enough to fill up a disk.
BizFlow Log Analysis Tips
It’s important to keep tabs on a variety of BizFlow logs, including:
- Individual components logs
- Archiving log
- SOAP logs
- ERA logs
- Form server logs
- Combined debug log
- Application server (Tomcat, etc.) logs
Each component of the BizFlow application generates its own log file. These individual component logs contain both performance and error information. They should be monitored continuously, as some errors only present themselves during heavy loads.
Archiving logs should be monitored to ensure that finished processes are being written to the archive tables and scrubbed from the active tables.
SOAP logs indicate problems with inter-system component communication and contain valuable profiling information. This profiling information can be turned into time-based metrics and used to predict issues and perform capacity planning analysis.
ERA logs track any activity on processes in response to events configured in ERA. Monitoring the ERA logs is critical for analysis of the ERA configuration and for troubleshooting any process problems.
Form server logs indicate any problems encountered in processing a form. Monitoring these logs can help you troubleshoot embedded code inside the form along with the actions the form takes.
Manual vs. Automated BizFlow Log Analysis
Without keeping close tabs on the BizFlow error logs, you can’t proactively manage your BizFlow environment. However, the sheer volume of data makes BizFlow log analysis a tedious and time-intensive proposition.
The typical manual approach to BizFlow log analysis involves opening a log file and searching for errors. However, in addition to being time-consuming, this method can generate numerous false positives – ranging from “normal” errors such as duplicate logons to inaccurate “timeout errors” for processes that completed successfully.
Another alternative to BizFlow log analysis is to narrow your manual search of the logs to look for only truly critical errors. There are two drawbacks to this approach: 1) you must first know what these truly critical errors are; 2) it’s virtually impossible to manually detect them amidst the noise of "normal" errors.
Human monitoring can’t possibly keep up with the volume of log messages generated by BizFlow – it’s like searching for a needle in a haystack. A better solution is to rely on a monitoring tool tailored for the BizFlow environment to filter out the noise and highlight real issues before they cause performance or availability problems.
System Shepherd®: A Platform for Proactive BizFlow Monitoring
Unlike generic application and database monitoring solutions, Absolute's System Shepherd® collects in-depth the BizFlow-specific metrics needed to facilitate meeting aggressive performance and availability SLAs.
System Shepherd continuously monitors BizFlow error logs for critical error messages and message frequency predictors, and alerts the appropriate person when it spots a problem or potential problem. System Shepherd also stores performance profiling information in a central repository, arming you with the data required to proactively analyze and tune your BizFlow environment on an ongoing basis.
