Scenes from a first timer’s trip to Datadog DASH

Quick summary: Datadog DASH 2026 brought together engineers and IT professionals from across industries to discuss observability, AI, and the challenge of managing increasingly complex software systems. Tricentis attended as a gold sponsor, where booth conversations revealed that performance validation remains a blind spot for teams on both sides of production.

The 2026 Datadog DASH conference was illuminating for a first timer. Attendees across vastly different industries — from hospitality to finance — gathered to learn about Datadog’s latest updates and sharpen their observability skills.

Representatives from Figma gave a talk about a system their engineers built to proactively monitor app performance and diagnose user-reported issues faster. Folks from JPMorganChase and Citizens Bank had a session on what it takes to innovate in regulated industries. Anthropic was represented in a discussion about deep reasoning with AI. And, of course, Tricentis’ Director of Customer Engineering, Bryan Cole, held a session on how some of the most valuable observability happens before production.

That’s a lot of ground to cover, and at first, it wasn’t totally clear to me how all these subjects were related to observability.

Bryan put it to me plainly: “We’re here talking about how our product serves people before, during, and after releases and how that makes them work faster,” he said. “It’s the one unifying theme here – everyone else is doing that, too.”

More from Bryan: How NeoLoad enables a continuous performance engineering loop

It’s not called DASH for nothing

Datadog CEO Olivier Pomel started the DASH keynote discussing the curve of productivity, which looks more like a vertical line all the time. With that productivity, he said, comes a good deal of complexity.

“We can now build technology in weeks or even days that would have been science fiction five or 10 years ago. That’s really amazing,” he said. “But it also means that we’re shipping apps and systems that we don’t fully understand with code that we mostly didn’t write, and there are a lot of components we don’t really have a good mental model for.”

That’s what Datadog does: helps its users detangle complex applications and identify pain points, so they don’t get in the way of productivity.

The keynote went on to announce some of Datadog’s innovations like Bits Investigation, an AI SRE agent, and Bits Release, a system for intelligent releases. Much of what the company announced was infused with AI.

Software is an industry at the center of a complicated technological time. AI enables developers to work faster, and when software developers can build a tech company’s product faster, the whole company works faster. That leaves plenty to detangle.

At the booth and on the stage: Putting performance in everyone’s hands

Tricentis had a booth right at the center of things — next to Google Cloud but not close to the popcorn. In order to earn a coveted mousepad, attendees had to get a “passport stamp” from a handful of booths, including ours.

This meant they had to talk to us.

A theme emerged from those conversations: Many software engineers and IT professionals either don’t know how their applications are load tested before release, or they don’t know if it happens at all.

That makes some sense. As Bryan put it during his breakout session, software engineers are constantly changing things while site reliability engineers and IT professionals are looking at the working application in production yelling, “don’t change anything!”

Performance validation is like a bridge between the essential SWE and SRE roles. It keeps both sides from having to make emergency fixes, but it’s rare that anyone crosses the bridge.

I found myself explaining it in the context of Black Friday. Because nobody wants their most profitable day of the year to be marred by an overloaded website.

More: How to handle traffic spikes when any day could be a holiday

As AI forces organizations and teams to speed up release cycles, performance validation will have to be engrained into the full software development lifecycle.

For SWEs, that looks like doing more than a unit test to ensure that your bit of code won’t cause a larger break in the system when you push it to production. For SREs, that means monitoring real user behavior and relying on continuous performance testing to avoid an expensive and time-consuming crash.

That’s the vision behind every update to Tricentis NeoLoad. From Augmented Analysis to Model Context Protocol (MCP) to Agentic Performance Testing and beyond, we’re adding features and functionalities that surface performance issues long before site outages or release delays.

Or, as Bryan likes to say, “You should know your code works before it’s in production, not just after.”

Dive deeper: Early insights, better outcomes: Shifting left with observability

Conferences: It’s worth coming out from behind the keyboard

Those of us who spend most of our day behind a keyboard and screen may find an event attended by thousands of people to be intimidating. Writers, like IT professionals, are stereotypically reclusive.

But as the novelty of AI begins to wane and organizations set about the difficult work of implementing and expanding it, events like Datadog inject some much-needed inspiration into our daily work.

Hearing from real people about how they’re using AI in their work and learning from some of software’s brightest minds gives us perspective on what’s possible when we look up from our desks and get in the room with colleagues.

Connection is necessary for innovation — we need to push each other to do better.

That’s also why Tricentis is hosting its 2026 Transform event in Dallas this year. We’ll be discussing AI, innovation, and the future of software testing.

See you there.