Dynatrace completed its acquisition of AI-observability company Arize on October 1, 2026. The strategic bet is that teams need to understand not only whether the software and infrastructure supporting an AI agent are healthy, but also whether the agent chose the right actions and completed its task. Dynatrace aims to connect Arize’s AI tracing and evaluation workflows with its broader application and infrastructure observability; that shared product experience remains a roadmap, not a completed integration.
Why Dynatrace bought Arize
Dynatrace announced the deal on August 13, 2026, as a cash-and-stock transaction valued at $915 million, then announced its completion on October 1. The company describes Arize as an AI observability and evaluation platform designed to support continual learning in agents. The rationale is to pair AI-specific evaluation with Dynatrace’s capabilities for monitoring performance, cost, and reliability across software systems.
As an Amazon Associate I earn from qualifying purchases.
The gap is practical: infrastructure can be functioning normally while an agent gives a poor answer, calls the wrong tool, or never finishes its assigned task. Diagnosing that failure may require following the agent’s behavior and model interactions alongside the APIs, services, and infrastructure it depends on. Observability for agents therefore has to connect two kinds of evidence rather than treating a healthy server as proof that the AI feature works.
What “billions of traces” means for AI agents
In an October 1 interview with The New Stack, Arize co-founder and chief product officer Aparna Dhinakaran said, “No human wants to go look at billions of traces.” Her point was that the volume of telemetry can make manual inspection impractical, creating a role for agents to interpret data and potentially act on what they find. The phrase belongs to Dhinakaran’s interview, not Dynatrace’s acquisition announcement.
#1 Best Overall
For an AI application, a useful trace may need to show more than a request’s duration or an error code. Teams may need to inspect model calls, retrieval, context, tool use, and the agent’s trajectory through a task, then evaluate whether the result was acceptable. Connecting that record to application and infrastructure telemetry can help explain whether a poor outcome came from agent behavior, a dependency, or the system around it.
How the platforms complement one another
| Dimension | Arize capabilities described by Dynatrace | Dynatrace capabilities described in the acquisition materials |
|---|---|---|
| Primary focus | AI application and agent traces, trajectories, evaluations, and debugging workflows | Performance, cost, and reliability across applications, services, infrastructure, user experiences, and business processes |
| Typical workflow | Inspect behavior, investigate failures, curate datasets, compare experiments, and iterate | Monitor production systems and understand operational context and downstream impact |
| Lifecycle role | Development and production workflows through Phoenix and managed platform Arize AX | End-to-end software and infrastructure observability, with an intended connection to AI evaluation over time |
Dynatrace identifies Phoenix as an open-source project and Arize AX as a managed platform for development and production workflows. Its stated direction is to connect these AI workflows with the wider observability stack over time. The companies have not described that future shared roadmap as a single, already-unified product.
Rank #2
What teams can do with Arize workflows
The described Phoenix workflow follows an iterative loop rather than a one-time production check:
The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →- Trace an application or agent and inspect its trajectory, including the steps it took toward a result.
- Run evaluations to assess behavior and investigate failures that need explanation.
- Curate datasets and compare experiments to understand the effect of changes.
- Use those findings to iterate on the system and continue evaluating it in development and production workflows.
When connected with broader observability, the intended benefit is to relate those AI-specific findings to the software and infrastructure context behind them. The acquisition announcement describes the strategic fit, while the integration of the workflows is a longer-term ambition.
Rank #3
OpenTelemetry and instrumentation
Dynatrace says OpenTelemetry formally accepted a code grant of Arize’s OpenInference GenAI instrumentation in June 2026. Incorporation into OpenTelemetry’s GenAI instrumentation project is incremental, according to Dynatrace; OpenInference remains an open, OpenTelemetry-compatible project. That is a standards and instrumentation development, not evidence that every product integration is already complete.
What the survey says about agent visibility
Dynatrace’s 2026 survey report covers 919 senior leaders globally and gives a margin of error of ±3.2 percentage points at a 95% confidence level. The figures below are from Dynatrace’s survey, not independent industry-wide measurements.
| Finding | Dynatrace 2026 survey result |
|---|---|
| Organizations with limited real-time visibility to trace and troubleshoot agent behavior | 42% |
| Organizations relying on manual methods to review communication flows among agents | 44% |
| Use observability during agentic AI development | 54% |
| Use observability during agentic AI implementation | 69% |
| Use observability during agentic AI operationalization | 57% |
| Report recording comprehensive logs and traces as a measure to validate agent decisions | 27% |
The figures suggest that use of observability varies by stage and that substantial visibility and validation gaps remain in the surveyed organizations. They describe survey respondents’ reported practices; they do not establish how every organization deploys agents.
Recommended Free Tools
Can agents analyze telemetry and fix software problems?
Dhinakaran told The New Stack that Arize’s Signal agent reviews traces from the company’s Alyx assistant, surfaces recurring issues, and opens pull requests. She said roughly 65–70% of those pull requests were accepted. This is her account of Arize’s internal experience, not an independently verified benchmark or a result established for other organizations.
The example illustrates the larger idea behind the headline: agents may help interpret telemetry at a scale that is difficult for people to review manually, and may propose changes based on recurring patterns. It does not mean that every trace should trigger an automated fix. Teams still need ways to evaluate behavior, inspect failures, and decide what changes are safe to accept.
What is completed—and what remains a roadmap
The acquisition is complete, and Dynatrace has explained the complementary roles it sees in Arize’s AI evaluation and tracing workflows and its own end-to-end observability. The next step described by the company is for the teams to shape a shared roadmap and connect workflows over time. Until that integration is delivered, readers should distinguish the combined strategic direction from capabilities available today as one unified product.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches
