Plumbr has decided to shift its flagship Plumbr product from a memory-leak tool to a JVM monitoring solution to broaden its appeal, according to company CEO Priit Potter.
Originally, Plumbr focused on automatically detecting root causes of memory leaks, and directing developers to the offending lines in their Java source code causing that problem. However, if a developer doesn’t have or doesn’t realise they have a memory leak, they might not use the tool. In practice, this means that when you realise you have a situation occurring in your environment you use it, otherwise you put it back on the shelf till the next time. Plumbr wanted to be a bigger part of the production scene, so decided they should evolve their product from a tool used to fix a specific problem, to a monitoring solution.
Since Plumbr v4.4, monitoring features added into the tool have included detection of thread contention, inefficient GC behaviour monitoring and historical data of your JVM to turn Plumbr from a ‘use when needed’ model to a ‘keep on’ solution.
The monitoring market already contains many good players and tools, several of which are very good at measuring and explaining the impact of performance issues. They can also explain the number of users affected and quantify the impact. Many of these tools are also good at localizing the misbehaving node in the infrastructure, so that the slow user experience can be traced down to “something bad” happening inside a particular JVM.
InfoQ talked with Plumbr co-founder and CEO Priit Potter about the change in direction of the Plumbr tool:
InfoQ: What were your reasons to switch from a memory leak solution to a monitoring solution?
Priit Potter: The main reason was that we did not want Plumbr to become a shelfware product that is only used when troubleshooting a specific problem. Plumbr technology offers most value in production environments, but you don't (want to) use troubleshooting tools in production. We found a way for our core technology to constantly deliver value to our users, and that's what led us to developing a monitoring solution.
InfoQ: Has Plumbr completed its switch to a monitoring tool, or is there more to come?
PP: There is indeed more to come. In addition to knowing when your systems are broken, a good monitoring tool is also able to confirm that everything is working smoothly. The first interesting results in this dimension will be launched in Q1-Q2 next year. We also have several other problem domains in research phase as eventually Plumbr will become the tool that catches all performance issues affecting the stability of the JVM.
InfoQ: How hard was the decision to change the shape of Plumbr, given it was already a successful tool?
PP: The more we learned about the monitoring domain, the larger and scarier the challenge looked, so indeed, it was not an easy decision. Nevertheless, seeing the way our value proposition was crippled when used only in development and test environments, it became an obvious choice. We do believe it was a justified change.
InfoQ: Given there are many other monitoring solutions on the market already, what is going to make Plumbr stand out?
PP: Plumbr is the only JVM monitoring tool that automatically detects the root causes of performance issues. It doesn't tell you that your Java app is running out of memory. It gives you the 10 lines of new code you need to copy-paste to your code base to make the issue go away.