Wrapping Up

This area, called DevOps, could be a book of its own. It’s easy to skip these important lessons as you adopt Elixir, but a little time spent with these tools can save you considerable effort as your system starts to grow. Erlang has a wonderful reliability and scalability history, owed in part to strong investments in DevOps.

In this chapter, we took the next step in your journey toward preparing your system for production. Our logging discussions showed you how to make the best out of the Elixir Logger, limiting logging to warnings and errors. We took a step-by-step tour through log customization, including custom formatters and back ends.

Then, we moved into tracing tools that can give you a good overview of what’s happening in your production system when things go wrong. You learned to use the break! debugger and SASL reports. You saw how tracers such as dbg are great fits when debugging concurrent systems with message queues.

Finally, we concluded the chapter by taking a look at three advanced tools to use when logging and tracing are not enough. Crash dumps with Observer give you a graphical view of a crash dump, runtime_tools provides good tools for runtime debugging and analysis, and the recon library offers a series of tools to simplify and analyze the many sources of production data.

With this wrap up, we’re through with this topic and this book. We hope you’ve enjoyed learning about the language, community, and ecosystem as much as we’ve enjoyed sharing it with you. You probably have an increased appreciation for those who’ve put in the thankless hours into building tools, writing prose, and blazing new ground with exciting applications. If so, consider giving back to the community that is supporting all of us, whether it’s a simple Stack Overflow post or a grand open source project of your own based on Elixir.

May your adoption story be as rewarding as ours was. Happy coding!