I’m not a car guy. I can drive them, but I don’t know the first thing about fixing them, and not much about maintaining them. My son, on the other hand – who’s getting a degree in mechanical engineering – is most definitely a car guy. He’s learning a lot about cars, and he’s teaching me a lot about what I’ve been doing wrong, even with simple things like maintenance.
I was thinking a lot about that as chapter 2 of my O’Reilly book, titled “Wrangling Telemetry” was being written. What book, you ask?
A long, long time ago (it was April 21) in a blog post far, far away (you can find it here) I wrote about the O’Reilly book that Avi Freedman and I are writing about Network Observability.
The book title is “Practical Guide to Modern Networking Telemetry”. The raw, “lightly edited”, somewhat unexpurgated version of chapter 2 is available today.
You can find it here: https://www.kentik.com/go/ebook/oreilly-guide-to-modern-networking-telemetry/.
I will briefly point out that it’s REALLY worth your time to download this copy because the final version will be about 120-150 pages (those O’Reilly editors don’t work for free!), but the raw version clocks in at 82 pages, and that’s just up to the end of chapter 2. So you really are getting a whole lot of extra content right now – content that will be lost to what looks to be a machete-and-chainsaw editing process later this year.
But if those 80+ pages aren’t enough for you – if you want a little more behind-the-scenes extra content to close our your week – I wanted to expand on just one idea that I throw out there in the first couple of pages of chapter 2.
On page 42 I wrote:
“Yes, it’s true that vendor solutions do a lot of heavy lifting behind the scenes. But this book is about exposing and elaborating on what that behind-the-scenes activity is. Part of the reason is so you can be a more informed customer. Knowing what happens inside the black box of observability solutions, HOW you get from “data comes in the left side” to “pretty charts come out the right side” allows you to make educated decisions about everything from OpenTelemetry methodology to aggregation options to pipeline management.
No matter how well the amazing folks at O’Reilly edit down chapter 2, the question remains: “Why should I read an entire chapter about how the observability sausage is made? Why do I care about enrichment, normalization, pipelines, data storage, and the rest. Like Avi and Leon SAY, RIGHT THERE IN THE BOOK, the vendor handles that shi…tuff.”
Let’s go back to my car analogy. It seems I’ve been spending a lot – too much – on oil changes. I thought that natural oil was better than synthetic. It was also significantly cheaper. With his experience on two different teams that compete in Baja SAE he’s learned a lot. And – without getting into the nitty gritty (because I still don’t understand all of it, and this is a blog about observability not oil) the synthetic oil is better for the cars we have, and will last longer, and overall work better.
That’s information I didn’t have because I didn’t choose to think about how my car operated on more than a surface level. I was content to let the vendor (the quicky-oil-change place) make my choice for me, and to make decisions based on price rather than effectiveness.
I find – especially when it comes to monitoring and observability – that IT practitioners can fall into the same trap. We already have too much on our plate. We already carry a huge cognitive load. Learning about the internal process of a tool that – let’s be brutally honest for a moment – is often seen as little more than something we do to satisfy an audit doesn’t seem like a good use of our time.
But – like that oil change analogy – it can have massive long-term impact. Understanding how your monitoring tools “wrangle” the data can
- help you avoid cost overruns
- understand why data doesn’t always match (what you expect or what other tools are telling you)
- use your tools both effectively and for the right use cases
- avoid wasting money replacing tools “that suck” when in fact they don’t suck, you’ve just been using them incorrectly.
- and more.
With all of that said, I don’t believe that knowing how to wrangle your telemetry is as useless as some people think. With the rise of Open Telemetry, those pipelines – and all the techniques to normalize, filter, sample, and transport data that they perform – is becoming more important than ever to understand.
So consider this chapter – in it’s full, lightly edited, unredacted form – to be Avi and my gift to you to become a better user of your monitoring tools.
Once again, you can find the book – all 80+ pages at this link: https://www.kentik.com/go/ebook/oreilly-guide-to-modern-networking-telemetry/.

0 Comments
Log in to join the conversation.No comments yet. Be the first to share your thoughts.