Dynatrace Expands into AI Observability
Dynatrace, a Massachusetts-based leader in application performance monitoring and optimization, has announced its expansion into the artificial intelligence space.
During its annual Perform conference, Dynatrace unveiled plans to enhance its core analytics platform with new features designed to track large AI models and their applications. This new offering, termed Dynatrace AI Observability, aims to equip enterprises with robust tools for monitoring generative AI systems as they gain traction.
The AI Observability Landscape
As organizations across various sectors increasingly adopt generative AI, they recognize its potential to enhance productivity and drive innovation. However, this technology also introduces risks, such as hallucinations, biases, and security vulnerabilities.
"We're committed to helping organizations navigate the challenges posed by generative AI, from security and transparency to reliability and cost management. Dynatrace is extending its observability and AI expertise to empower customers to embrace AI confidently, ensuring unparalleled insights into their generative AI-driven applications," stated Bernd Greifeneder, Dynatrace's CTO.
Capabilities of Dynatrace AI Observability
Generative AI is essential for business competitiveness, offering opportunities for improved efficiency and automation. However, it can also lead to high costs and inaccurate outputs if not closely monitored. Addressing these challenges requires vigilance to identify and resolve issues like model drift or data discrepancies. This is where Dynatrace AI Observability plays a crucial role.
How Dynatrace AI Observability Works
By integrating metrics, logs, traces, problem analytics, and root-cause information, Dynatrace's observability solution provides end-to-end monitoring of modern AI applications. It spans the entire AI stack—from infrastructure components such as Google TPUs and Nvidia GPUs to foundational models like GPT-4 and advanced frameworks used in modern architectures.
This comprehensive approach grants teams visibility into the AI application's lifecycle, allowing for the identification of performance bottlenecks and systemic issues. For instance, the solution can analyze infrastructure utilization—which encompasses temperature, memory, and process usage—thus optimizing resource consumption and operational costs.
Seamless Integration and Compliance
Dynatrace AI Observability also features integrations with cloud services and custom models, such as OpenAI and Azure services. These integrations ensure robust model monitoring while adhering to service-level agreement (SLA) performance metrics, including latency, availability, and error rates. The solution is equipped with Dynatrace's proprietary Davis AI engine, enhancing its ability to monitor AI application outputs and align with privacy and security regulations.
Availability and Future Outlook
Dynatrace is rolling out its AI observability solution to all customers starting today, with select companies like OneStream having received early access for testing. "To implement generative and predictive AI successfully, we rely on Dynatrace for its reliability and performance in supporting critical workloads," remarked Ryan Berry, SVP of engineering & architecture at OneStream.
As investments in generative AI continue to grow—projecting the market to reach $1.3 trillion by 2032—monitoring will become pivotal for organizations. However, Dynatrace is among numerous players in the AI observability market, competing with companies like Monte Carlo, Arize, and Datadog. The coming years will reveal how enterprises leverage solutions like Dynatrace AI Observability in this rapidly evolving landscape.
Ranking
