Book Review ~ Practical Monitoring: Effective Strategies for the Real World by Mike Julian
My very first book review; I bet you’re all so excited. I know I am!
Instead of giving a numerical score (what does that even mean?!), I’m going to break each book down with my overall impressions, key takeaways, and suggesting audiences I think will benefit most. The format of these reviews may change over time, but for now, let’s dive in to Practical Monitoring: Effective Strategies for the Real World.
Practical Monitoring: Effective Strategies for the Real World is a tool agnostic overview of monitoring principles and key metrics to monitor in several areas – business KPIs (key performance indicators), frontend, application, server, network and security.
The layout of the information Mike presented was well-chosen. Starting with the anti-patterns of monitoring (think the single pane fallacy, obsession with tools, alerting on unhelpful things like OS metrics, being too conservative in what you’re collecting) was really smart. Once you have a handle on everything you shouldn’t be doing, Mike’s lessons on what you should be doing are easy to grok.
Any good monitoring book will include a chapter on alert fatigue and on-call best practices. There’s always room for improvement wherever you work, and that wasn’t overlooked here. The chapter on the math of metrics was also a nice touch. Data you’re collecting is pretty useless if you don’t know how to interpret it.
I would’ve liked more information on monitoring microservice architecture… that could’ve been a whole chapter given that’s the direction a lot of companies are starting to explore. I also felt that the final chapter on conducting a monitoring assessment was too brief. Some final nuggets of wisdom would’ve been great here.
Start from the top down with your monitoring; determine what your business KPIs are, then translate those into technical things you can actually monitor and alert on.
Define what “working” means for your application, and monitor that. Sure, collect OS-level metrics like disk space, memory and CPU load, but don’t just alert on arbitrary values. Alert on rates of increase/decrease or unusual patterns.
There is no shame in buying monitoring tools, and in fact, this is usually the best option. The cost ends up being far less than paying a team of engineers to build and/or maintain a homegrown monitoring solution. On top of that, these tools are built by monitoring experts. Can you say that about your engineers?
Practical Monitoring is a great overview for engineers that are brand new to monitoring, or know that their current monitoring isn’t serving them, but they’re not sure where to start in improving it.
As this book isn’t super technical, and anything technical is explained in a clear way, I’d even venture to suggest this book for business owners. Monitoring is such an important, yet so overlooked area, and this book can go a long way in helping a company tie it’s bottom line to monitoring the health of its front-end, applications, and infrastructure.
If you’re experienced in the art of monitoring, this is probably not the book for you. And kudos to Mike Julian for mentioning this in the Preface section titled “Who Should Read This Book”.
Have you read Practical Monitoring? What did you think? Or are you curious to read it now?