Culture / Development / Monitoring / Sponsored

Why High-Performing DevOps Teams Thrive on Observability 

15 Jun 2021 4:00am, by

Consistent and rapid delivery of software releases and updates, and the ability to quickly fix code when things go south in minutes instead of in hours or even days are earmarks of a well-functioning DevOps team. Getting there depends on having several socio-techno factors in place, including the adoption of a proper observability framework and culture.

Proper observability also does not hinge as much on access to logs, metrics and tracing data, while important. Instead it allows for inferences and fixes to be made in a way that reflects the way high-performance teams are able to optimize their production environments. This was a key takeaway from o11ycon+hnycon, Honeycomb’s annual user conference.

Indeed, proper observability should allow for the analysis of application performance in an actionable way from the beginning to the end of production without overemphasis on logs, metrics and tracing capabilities.

This holistic approach to observability helps to determine if code is ready to deploy as well as its resiliency before and after deployment. It also helps to analyze customer user experience post-deployment, and ultimately, to determine whether applications are meeting service-level objectives (SLOs).

“That is the promise of observability: It’s that drive, that curiosity, that ability to understand not just how your applications are behaving today, but how to make your software perform better tomorrow,” Christine Yen, CEO and co-founder at Honeycomb, said during “The Future of Observability” keynote. “You look at the product that Honeycomb has built and, continues to build, and you’ll see that these correlates are manifested in the experience we craft for you.”

Honeycomb’s platform “under the hood” features a technical architecture that allows organizations to “capture data that describes your business, your customers, your configuration, in a way that makes sense to your engineers, who shouldn’t have to predict ahead of time which attributes will be important, alongside other attributes, so that you can build indexes on them,” Yen said.

Not to say metrics do not have their place in the process, since a proper observability platform should use logs and tracing capabilities to gather this information through queries on an as-needed basis. To that end, Honeycomb released during the conference Honeycomb Metrics, which offers a single interface for observability and metrics data. This removes the need to seek metrics data through event-based models, while using traces and logs in native format, without requiring users to convert time-series metrics into events.

During a demo “Introducing Honeycomb Metrics” given by Max Edmands, a staff product engineer at Honeycomb,  she showed how the interface offers a “collection of features for enterprise accounts that will give you and your team a window into what’s happening in your infrastructure.“ Honeycomb has always been a state-of-the-art tool for observing what your services are doing in production, but sometimes it’s important to go one level deeper and get some insight into what’s happening inside the infrastructure that your applications rely on,” Edmands said. “Wouldn’t it be great if we were also able to look at infrastructure metrics on this screen?”

Bug Find and Fixes

When code does go awry and things break, which, as everyone knows they do often, reliance on observability for debugging is critical. For instance, Erwin van der Koogh, product manager at Cloudflare, during a talk entitled “Killing Your Staging Environment,” said good observability for debugging is “magic fairy dust.”

“This is the closest thing I can find to magic fairy dust because fixing bugs becomes almost trivial, like it feels like cheating, basically, like if you have that good observability — I can thank Honeycomb for that —locating the issue is easy,” he said. “Again because it’s production data you’re looking at and you have the exact steps that happened to [create the bug], you have that full capability to make a change, push it to production, test it, and if it’s not that, just go try something else. You have this ability to iterate in production, fully confident that you’re not affecting anything else [to the application code] or to this particular customer that is having this particular issue.”

During a keynote, Rich Anakor, chief solutions architect at Vanguard, described how Honeycomb’s observability tools, as well as OpenTelemetry, a Cloud Native Computing Foundation (CNCF)  vendor-neutral framework that provides interoperability among different tools, played an instrumental role in the financial services firm’s digital transformation.

During the migration process to multicloud environments from a legacy data center infrastructure, Anakor described how a Vanguard team was moving data to a new repository in the cloud and the challenges they faced.

“The team wanted to know all the dependencies that were involved,” he said. “They wanted to know all the user actions and how they mapped back to these backend stock procedures. They had been studying spreadsheets for months looking at code, involving really smart people engaged in trying to solve this problem, but because this was considered a legacy application, they did not think we could help this application, which also had dependencies with other workloads in the private cloud. We said, ‘Let’s try this out with OpenTelemetry and Honeycomb,’ and we were able to answer these questions within minutes — minutes! That was key.”

Anakor’s testimonial about how Vanguard used Honeycomb to change developer workflows, and to achieve a much better understanding of what was happening in production illustrated how observability offers real, tangible benefits for a DevOps team’s entire workflow. “It’s pretty exciting to see people building their careers on making their lives better and their co-workers better like this [in a way] that we’ve always kind of looked at as a dream,” Charity Majors, CTO and co-founder of Honeycomb, said during a keynote.

A newsletter digest of the week’s most important stories & analyses.