TL;DR — Key Takeaways
– Splunk plans to release an open source AI model on Hugging Face that is specifically trained to analyze and reason across log data.
– Splunk is also developing a Universal Collector, expected in beta in 2027, to simplify telemetry collection using OpenTelemetry.
– The new log-focused model complements Cisco Time Series Model 1.0, which was designed to analyze metrics.
Splunk is gearing up to make an artificial intelligence (AI) model for analyzing log data available on Hugging Face under an open source license.
Additionally, Splunk at its .conf26 conference this week revealed it is developing a Universal Collector, expected to be available in beta in 2027, to streamline collection of all types of telemetry data using an instance of OpenTelemetry.
Raja Mukhopadhyay, vice president of observability cloud for the Splunk arm of Cisco, said a large language model (LLM) that has been trained to reason across log data will make it possible for DevOps teams to observe AI applications and agents at scale. The Universal Collector, meanwhile, lays the foundation for unifying DevOps, IT operations and security operations by making it simpler to correlate events without having to deploy separate repositories for different types of telemetry data.
Previously, Splunk has made available an open source Cisco Time Series Model 1.0 on Hugging Face that has been specifically trained to analyze metrics. Both LLMs are needed because general-purpose AI models were trained using text, code and video. Metrics, in contrast, are a type of numerical data that requires an LLM that has been specifically trained to consume that data, said Mukhopadhyay.
Logs, and by extension, traces, could, of course, be analyzed using a general-purpose model, but the volume of that data would overwhelm the amount of memory allocated to a context window. The second AI model for log data that Splunk has developed is, in contrast, trained to reason across that data more efficiently, noted Mukhopadhyay.
As DevOps teams employ AI agents within workflows, the amount of telemetry data being generated is exploding. As such, the need to find a way to analyze that data in a way that enables DevOps teams to understand the behavior of those AI agents at this juncture has become all too apparent, noted Mukhopadhyay.
Less clear is what DevOps teams will be observing in the era of agentic engineering. Most AI agents today are generating code using programming languages that were designed for humans. However, it’s probable that at some point AI agents will generate more efficient code using some type of machine language that humans will not be able to read. The assumption is that other AI agents will be deployed to validate that code is working as intended.
In the meantime, however, DevOps teams should expect the number of applications running in production environments to continue to increase exponentially. The challenge then becomes how to manage those applications as they are continuously updated by what will soon become swarms of AI agents. Rather than being in the middle of workflows that will be executing at machine speed, software engineers will instead need to evolve into managers of fleets of AI agents that have been assigned specific tasks to autonomously perform.
Hopefully, the end result will be not just more software being built and deployed faster than ever but, one day, higher quality applications. After all, using AI to deploy more flawed software faster than ever is more than likely going to turn out to be more trouble than it’s worth.
Frequently Asked Questions
What AI model is Splunk developing?
Splunk is developing a specialized large language model designed to analyze and reason across log data.
Where will Splunk make the model available?
Splunk plans to make the model available on Hugging Face under an open source license.
What is Splunk’s Universal Collector?
The Universal Collector is a planned telemetry collection tool based on OpenTelemetry that is intended to streamline the collection of logs, metrics and traces.
