At its annual Perform conference, Dynatrace unveiled its foray into the artificial intelligence space with the introduction of Dynatrace AI Observability. This new offering aims to empower enterprises to monitor and optimize both large AI models and the applications powered by them. With the increasing adoption of generative AI systems, Dynatrace AI Observability fills a crucial need for enterprises seeking to closely track the performance and reliability of these advanced technologies.
In a statement, Bernd Greifeneder, CTO at Dynatrace, emphasized the transformative potential of generative AI while acknowledging the challenges it poses in terms of security, transparency, reliability, user experience, and cost management. Dynatrace AI Observability is designed to address these challenges comprehensively, providing organizations with the insights needed to confidently deploy and manage generative AI-driven applications.
Key Features and Functionality
Dynatrace AI Observability leverages the company's extensive analytics platform to offer end-to-end monitoring of the entire AI stack behind modern applications. This includes monitoring infrastructure layers such as Google TPUs and Nvidia GPUs, foundational models like GPT-4, semantic caches, vector databases, and orchestration frameworks like LangChain.
By consolidating metrics, logs, traces, problem analytics, and root-cause information, the observability solution provides teams with a holistic view of the AI application lifecycle. This enables them to identify performance bottlenecks, root causes of issues, and areas for optimization.
The solution integrates with various cloud services and custom models, including OpenAI, Amazon Translate, Amazon Textract, Azure Computer Vision, and Azure Custom Vision. This integration facilitates robust model monitoring, including tracking service-level agreement (SLA) performance metrics such as token consumption, latency, availability, response time, and error count.
Dynatrace AI Observability is also equipped with the company's proprietary Davis AI engine, enabling precise tracing of AI application outputs. This ensures compliance with privacy and security regulations, as well as governance standards.
Availability and Early Adoption
Dynatrace has made AI Observability available to all its customers, with early access provided to select companies like OneStream for testing purposes. Ryan Berry, SVP of engineering & architecture at OneStream, praised Dynatrace for its leadership in AI and observability, highlighting the importance of reliable and performant services for supporting critical workloads.
As enterprises continue to invest in generative AI technologies, the adoption of observability solutions like Dynatrace AI Observability is expected to grow. However, Dynatrace faces competition in the AI observability space from other players such as Monte Carlo, Arize, Context AI, Weights & Biases, and Datadog.
Dynatrace's entry into the AI observability market underscores the growing importance of monitoring and optimizing generative AI applications, positioning the company as a key player in enabling enterprises to harness the full potential of AI while mitigating associated risks.
