Supporting digital transformation across 36 campuses

UNESP (Universidade Estadual Paulista) is one of Brazil’s largest public research universities, serving more than 55,000 students across 36 campuses throughout the state of São Paulo. Supporting that mission requires a highly distributed technology environment that powers everything from student enrollment and learning platforms to research computing, administrative systems, and cybersecurity operations.

The university’s central IT division, CTInf, manages approximately 200 institutional applications running across hybrid infrastructure that spans on-premises data centers, Kubernetes environments, cloud platforms, and one of Brazil’s largest academic networks.

As UNESP accelerated its digital transformation initiatives, maintaining visibility across that environment became increasingly important. Infrastructure, application, and security teams relied on multiple monitoring tools, each providing only a partial view of the university’s systems. Correlating issues across environments was often difficult, and in many cases users reported problems before IT teams were aware of them. “Before Datadog, each team had visibility into only part of the environment. Bringing metrics, traces, and logs together gave us the context we needed to understand issues across the university as a whole,” says Ney Lemke.

To support reliable digital services for students, faculty, and researchers, UNESP needed a more unified approach.

Seeing the university through a unified view

After evaluating observability platforms and consulting with peer institutions in Brazil and internationally, UNESP selected Datadog as the core of its observability strategy.

By adopting Infrastructure Monitoring, APM, Log Management, Database Monitoring, Code Security, and App and API Protection, the university created a single view across its hybrid environment. Teams could now correlate metrics, traces, logs, infrastructure events, and security insights within one platform. “The ability to correlate infrastructure, applications, and security data in a single platform fundamentally changed how we troubleshoot problems. We spend less time searching for answers and more time resolving issues,” tells Carlos Coletti — IT Manager - FUNDUNESP.

That visibility quickly proved its value. During a new student enrollment period, the university’s Access Portal experienced performance issues affecting users across multiple campuses. Datadog traces revealed that the bottleneck was not in the application itself, but in the web server responsible for distributing requests. Scaling that component resolved the issue and restored performance.

“Before Datadog, each team had visibility into only part of the environment. Bringing metrics, traces, and logs together gave us the context we needed to understand issues across the university as a whole.”

In another case, performance issues during peak demand for university restaurant voucher purchases were traced to an ingress proxy configuration limit rather than application or database performance. Datadog helped the team identify the true root cause and restore service quickly.

With a unified observability platform, UNESP could move beyond assumptions and focus directly on solving problems.

Turning observability into a conversation

As observability adoption expanded, CTInf identified another opportunity: making Datadog insights accessible to more teams.

Although Datadog provided comprehensive visibility into logs, traces, and metrics, extracting insights still required users to navigate dashboards, understand query syntax, and manually correlate information across multiple views.

To address that challenge, the team developed cAIo, an internal observability assistant powered by Datadog MCP Server. The goal was to make Datadog easier to use.

Datadog MCP Server became the engine behind the experience. When a developer asks a question in natural language, MCP translates that request into Datadog queries, retrieves the relevant traces, logs, and metrics, and returns a concise explanation based on real telemetry. “Datadog MCP Server fundamentally changed how people interact with observability data. Instead of teaching every developer how to navigate dashboards and write queries, we can let them ask a question and get an answer directly from Datadog.” says Carlos Coletti.

What began as a tool for SRE and DevOps teams quickly expanded to development teams across the university. The result has been an 80% reduction in investigation time, with many common observability tasks dropping from as much as five minutes to under one minute.

Using Datadog traces surfaced through cAIo, the team identified a 13-second trace in which 99% of execution time was spent waiting on an upstream dependency, immediately directing engineers to the correct service for optimization.

“Datadog MCP Server fundamentally changed how people interact with observability data. Instead of teaching every developer how to navigate dashboards and write queries, we can let them ask a question and get an answer directly from Datadog.”

In another case, cAIo used Datadog logs and APM data to identify a spike in java.lang.NullPointerException errors, locate the failing endpoint, and surface the relevant stack trace in under 10 minutes.

It also proactively detected a memory leak using Datadog Kubernetes telemetry before the issue could trigger a production outage.

By combining Datadog observability with natural-language access through MCP, UNESP transformed observability from a specialized discipline into a capability available to a much broader set of teams.

Advancing digital transformation across higher education

Today, UNESP’s CTInf team uses Datadog to support the university’s mission of delivering reliable digital services to students, faculty, researchers, and administrative staff. With unified visibility across its distributed environment, the team is able to maintain its target of greater than 99.7% system uptime, accelerate troubleshooting, and strengthen collaboration across infrastructure, application, and security teams.

CTInf has also turned observability into a powerful communication tool for university leadership. Through real-time dashboards and service level agreement (SLA) reporting, the team gives governance bodies clear visibility into service performance, helping them communicate operational health, demonstrate the impact of their work, and make data-driven decisions about future infrastructure investments. Lemke explains, “Datadog has become more than a monitoring platform for us. It’s a foundation for digital transformation, helping us deliver better services while giving leadership real-time visibility into the health of our environment.”

Looking ahead, UNESP plans to expand observability across AI-powered applications, deepen monitoring of its academic network infrastructure, continue evolving cAIo through Datadog’s AI capabilities and MCP integrations, and support the implementation of its upcoming cybersecurity policy.

“Datadog has become more than a monitoring platform for us. It's a foundation for digital transformation, helping us deliver better services while giving leadership real-time visibility into the health of our environment.”

For UNESP, observability has become a foundation for delivering reliable digital experiences, supporting research and innovation, and advancing digital transformation across one of Brazil’s largest public university systems.