Datadog sees data observability join DevOps, adds Metaplane
Datadog has customers already bringing data science under DevOps and, with its latest acquisition, aims to create a unified tool set for the newly converged combo.
Digital transformation and AI have some enterprises consolidating data science and DevOps teams, Datadog officials said. The vendor plans to offer them data observability baked into its platform later this year, based on its acquisition of Metaplane.
The data observability startup, founded in Boston in 2019, uses machine learning to automatically alert for issues with data quality and delivery in big data warehouse systems such as Snowflake. The two companies share joint customers but declined to say how many. Metaplane competitors in data observability include Acceldata, Monte Carlo and Soda Data.
Metaplane's founder, Kevin Hu, and his engineering team will join Datadog and rewrite the company's software to run on Datadog's observability platform. Financial terms of the deal were not disclosed.
Datadog chose to acquire Metaplane because it's both willing and able to rewrite its tools to run on Datadog, according to Michael Whetten, vice president of product at Datadog. For now, Metaplane will continue to support existing customers as Metaplane by Datadog, but Whetten said its IP will be ported to the Datadog platform within the year. After that, customers will need to move to Datadog to continue to use Metaplane data observability.
The deal came about because Datadog has already seen some customers bring DataOps and DevOps teams together and begin to ask for unified tools in the last year, Whetten said.
"We saw that data engineers and data teams were starting to consolidate budgets under the CTO and the application team," Whetten said in an interview with Informa TechTarget. "There was an indication that these things were moving out of just business intelligence, which is still critical, toward real-time operations at scale."
Metaplane also had customers asking it to cover more of the IT infrastructure used for data processing, including data pipeline tools such as Kafka, according to Hu.
"Data comes from somewhere -- a system, or input by a human, and unfortunately, Metaplane and the current cohort of data observability customers just don't have that visibility," Hu said. "We're not connected to the Kafka stream. We're not connected to the upstream database or to upstream software systems. So when our customers naturally ask the 'What happened?' question, we just simply don't have an answer."
Ultimately, porting Metaplane into Datadog will add more value for users, he said.
"We don't know exactly how it's going to look or what form it's going to take, but now we have data to answer that question of exactly what happened," Hu said. "And if you're writing software or you're on a DevOps team, you can answer the reverse question of 'What will happen if I make this change?' so it's the full picture."
A harbinger of M&A to come?
Industry analysts haven't yet seen consolidation between data science and DevOps teams en masse, but don't discount the possibility that it could become a trend as data hygiene and DataOps techniques play a bigger role in the development and maintenance of generative AI and agentic AI applications.
"This might be the first large acquisition like this, merging the tools that the data teams use and the tools that the IT teams use … which otherwise have been in different worlds," said Gregg Siegfried, an analyst at Gartner. "I've kind of been expecting, as the interest in AI has grown, that those worlds being separate is not necessarily going to be permanent."
It could be an indication that larger things are afoot. You might see other data observability companies get acquired by observability vendors.
Gregg SiegfriedAnalyst, Gartner
Datadog rival Dynatrace has added data observability features to its platform with relatively little fanfare over the last 18 months, Siegfried said.
"It could be an indication that larger things are afoot," he said. "You might see other data observability companies get acquired by observability vendors, if that becomes a competitive [differentiator]."
Whether data science and DevOps teams consolidate, platform engineers and SREs must consider data warehouse and analytics systems when troubleshooting machine learning or generative AI-driven apps. Poor data quality can also lead to poor AI automation results.
"In the era of AI, data hygiene has become very important," said Chris Condo, an analyst at Forrester Research. "Large language models are now critical components of modern application architectures, so this feels like a natural extension into that area of software development."
Business intelligence and business observability have similarly grown in their significance to enterprises in the last decade, as has observability in general, prompting the need for high-scale data management tools such as data warehouses, Whetten said.
"AI obviously is taking things to another level of scale, [but] even just moving to the cloud and adopting something like a container orchestration tool produces a lot more telemetry data," he said. "So, on the big data side, if you look at what people are putting into data warehouses … it's because user engagement and the types of usability and access logs that go into a data warehouse are also growing."
Beth Pariseau, a senior news writer for Informa TechTarget, is an award-winning veteran of IT journalism covering DevOps. Have a tip? Email her or reach out @PariseauTT.
