One of the things about OpenTelemetry that’s easy to miss if you’re not spending the whole day in the ins and outs of the project is just how much stuff it can do—but that’s what I’m here for! Today, I want to go through the project and give you a guide to the various parts of OpenTelemetry, how mature they are, and what you can expect over the next six months or so. I ranked these elements by relative maturity across the entire project.

In 2016, we at Honeycomb first borrowed the term “observability” from the wikipedia entry for control systems observability, where it is a measure of your ability to understand internal system states just by observing its outputs. We then spent a couple of years trying to work out how that definition might apply to software systems. Many twitter threads, podcasts, blog posts, and lengthy laundry lists of technical criteria emerged from that work, including a whole ass book.

Engineering has come a long way since the days of delivering discrete, point-in-time products that were often packaged on a CD and shipped to customers. The days of physical media and long development cycles are long gone. The advent of cloud computing and the rise of Software-as-a-Service (SaaS) transformed the landscape, creating a new model of continuous development and service delivery. This shift has not only revolutionized how software is developed, but has also redefined the engineer’s role.

Sampling is a necessity for applications at scale. We at Honeycomb sample our data through the use of our Refinery tool, and we recommend that you do too. But how do you get started? Do you simply a set rate for all data and a handful of drop and keep rules, or is there more to it? What do these rules even mean, and how do you implement them? To answer these questions, let’s look at a rules file template that we use for customers when first trying out Refinery.

As part of our recent failure testing project, we ran into an interesting failure mode involving the OpenTelemetry SDK for Go. In this post, we’ll show you why our apps stopped sending telemetry for over 15 minutes and how we enabled keepalives to prevent this kind of failure from happening in the future.

Monitoring used to be so easy. Servers had names and lived down the hall, or across the street. If things weren’t working, you could turn them on and off again. Database filling up? Just throw another hard drive in there. Too many simultaneous requests? Rack another server and install a cache. Fast forward a couple decades, and things have gotten much more complicated.

Setting clear, measurable goals is essential for any successful team. However, aligning those goals with the technical work can be challenging in the fast-paced world of software engineering. Engineers might focus on reducing latency or improving uptime, while business leaders look at revenue and customer satisfaction. It gets tricky to track the impact between the two to justify when specific engineering initiatives are important, why, and how they impact the bottom line.