Let me start by defining the current problem-detection process in most enterprises. An issue arises in the production environment for a J2EE application running on WebLogic, information is captured via logs, and WebLogic server is rebooted. From this point, application developers and administrators are tasked with analysis of the information and finding the root cause.

If they are skilled and fortunate enough they will have all of the information. However in most situations this isn't the outcome. You rarely find the details needed for root-cause analysis from production failures. You frequently have to reproduce the problem in a test environment to be able to get the proper data to determine "root cause." You get the picture: all of this can require significant time and effort.

After facing these problems in the field for many years as a WebLogic consultant and finding available tools lacking in this area, I decided to do something about it. Arcturus (www.arcturustech.com) AutoPilot for WebLogic is the result of that decision.

AutoPilot addresses these issues with a unique approach. AutoPilot is an expert system-based tool with many person years of expert knowledge embedded within. AutoPilot automates extremely tedious tasks such as Proactive Monitoring, Instant Root Cause Analysis, and WebLogic Tuning, WebLogic environment review for best practices. AutoPilot results in significant savings in terms of man hours and hardware/software licenses.

AutoPilot proactively monitors WebLogic using its IntelliCheck Technology for patterns that can cause an outage and warns administrators before they actually do. AutoPilot doesn't require any instrumentation to discover unusual patterns. AutoPilot uses information already available in WebLogic and has extremely low overhead.

AutoPilot completely automates the process of root-cause analysis. AutoPilot automatically detects WebLogic failure conditions in production and auto-analyzes WebLogic state (thousands of matrices) to generate a root-cause analysis report instantly. Additionally, AutoPilot will generate an e-mail that not only tells you that WebLogic server failed, but that also tells you why.

AutoPilot is an all JAVA and J2EE solution that plugs right into and seamlessly coexists with BEA WebLogic Server and Console (see Figure 1). You can get an evaluation version of AutoPilot at http://support.arcturustech.com/downloadpage.do.

The installation process for AutoPilot is straightforward. The AutoPilot installation wizard installs AutoPilot in fewer than 10 minutes. You can obtain more details on AutoPilot installation at http://support.arcturustech.com/APHelp/installingautopilot.htm.

Although WebLogic does provide all JMX information, it is not in an easy-to-use format. AutoPilot provides an easy-to-use interface for browsing MBeans information. With AutoPilot explorer, views are available for exploring configuration and run-time MBeans properties. AutoPilot explorers make JMX information accessible directly within WebLogic Console, without causing you to have to browse through many JSP pages to obtain the same.

One of the features that I missed with WebLogic was the lack of persistence capabilities for monitoring data. You do get a nice chart of heap and queue size in WebLogic, but once the information falls off the screen it is gone forever. What that means is if you are not in front of the Console when something bad happens, you are out of luck. AutoPilot automatically persists key information about WebLogic state for historical analysis. It provides WebLogic Administrators with extremely easy-to-use data-mining capabilities. For example, you can start with a 30-day view of WebLogic server's performance and drill down to a particular time frame in a matter of a few clicks. AutoPilot Analyzers help you spot problems & trends within the WebLogic Server (see Figure 2).

AutoPilot allows you to create sensors on any JMX property. A sensor is itself a JMX property whose value is persisted every time AutoPilot takes the pulse of WebLogic Server. AutoPilot provides an easy-to-use interface to graphically analyze data accumulated by AutoPilot sensors.

One of the unique and very useful features of AutoPilot is Blackbox. AutoPilot Blackbox (Figure 5) allows you to take a snapshot of the WebLogic Server state at any point in time. Blackbox is the view of the internal state of WebLogic Server and the applications running on it. Blackbox saves the state of the server at the time when it was created. Blackbox data can be analyzed with an easy-to-use explorer view, similar to the one used by AutoPilot Config and Run-Time MBeans Explorers. You can create a Blackbox at any time just with the click of a button. It comes in handy when you are in the middle of an analysis and want to preserve the state.

From my experience, it is much easier to find the root cause of the problem if you see it happening while you are there. Unfortunately however, you can't have an expert sitting in front of WebLogic Console waiting for a problem to occur. AutoPilot solves this problem. A key feature of AutoPilot is the auto-detection of failure conditions in WebLogic Server. AutoPilot automatically detects when WebLogic Server gets into a bad state. Once such a condition is detected, AutoPilot automatically creates a Blackbox for WebLogic. In most cases once WebLogic gets into this bad state, administrators bounce the faulty instance of WebLogic Server quickly to bring the production environment back to the desired capacity. Hence even if an expert is available to analyze the server and find the cause for the issue, the person doesn't have enough time for troubleshooting. Blackbox provides these experts with the complete state of the server at that instance, even after the server is brought down. AutoPilot just doesn't stop there. It auto analyzes the state of the WebLogic Server to generate a "root cause" report that gets e-mailed to the administrator(s) automatically.

In this way experts get all of the information they need even though the incident may have happened in the middle of the night while they were catching up on their sleep. AutoPilot can analyze thousands of matrices almost instantly, versus an expert who would take significant time to analyze the same. AutoPilot drills down to transaction level in determining the root cause and identifies the culprit transaction(s). This is what makes AutoPilot for WebLogic an extremely useful tool.

The AutoPilot Detector that determines the root cause can also be manually executed to analyze the run-time state of WebLogic Servers at any point in time. It is extremely efficient and it's the least intrusive way of finding out the problematic patterns developing in your environment. AutoPilot detector doesn't make any changes to the environment and hence can be safely executed in a production environment.

AutoPilot adds sought-after self-detection capabilities to WebLogic 8.1. Using its IntelliCheck Technology, AutoPilot monitors and analyzes internal WebLogic state to detect conditions that can lead to an outage. Once such a condition is detected, AutoPilot raises alerts for administrators.

As a WebLogic consultant, every customer I visited asked me to review their environment and guide the on where there is room for improvement. It is a time-consuming task to certify that there are no issues with the environment and advise about best practices. With AutoPilot Advisor we have automated this task. AutoPilot Advisor analyzes your whole WebLogic environment and generates a report of recommendations. AutoPilot Advisor tells you where you are not leveraging your WebLogic right and guides you on how to resolve those issues. Advisor also guides by revealing best practices and where in your environment you can benefit from them. This is equivalent to having an expert evaluate your environment and provide feedback on how well you are leveraging your infrastructure. The AutoPilot advantage is that it applies many person years of expert knowledge at the touch of a button and produces results in a matter of minutes.

Many times you have a production outage and you work with BEA Support to find out that there is a patch (CR) available for that. Wouldn't it be good to know of these patches that are applicable to your environment before you actually encounter the issue in production? With AutoPilot you can. Advisor checks for all the applicable patches to your environment and provides you with details on each. It doesn't just blindly give you a list of patches that are available from BEA in newer service packs, it goes and checks for relevance. This way we are able to filter the list of patches to the ones you must have.