University of Iowa College of Public Health builds unified observability with Grafana Cloud

The University of Iowa is a leading public research institution, home to more than a dozen colleges and a wide range of federally and privately funded research programs. This includes the College of Public Health (CPH), which supports studies in areas such as cancer statistics, population health, and environmental health—each requiring reliable access to computing and storage resources.

Within the CPH, a three-person IT team manages the research systems that power these workloads. Their scope is intentionally broad: from servers and hypervisors to specialized research environments, these disparate resources are used by faculty and graduate researchers who depend on accurate data to inform grant decisions and scientific work.

The CPH was using Grafana OSS on premises to monitor these systems, but the constant maintenance was overwhelming the small IT team. That’s when they decided to migrate to  Grafana Cloud, establishing a reliable, scalable, and cost-efficient observability foundation that would help them keep the CHP’s workloads running much more efficiently. 

After reducing ingestion from more than 630,000 metric series to about 4,000, the team now has a clear, actionable view of every system they manage on premises and AWS— even lowering their observability costs by more than 98% along the way!

 With Granfana Cloud powering dashboards, alerting, and predictive insights, CPH’s small IT staff can support a wide range of research workloads with greater confidence, visibility, and efficiency. The college is even expanding into logs and cloud cost dashboards to deliver a true single pane of glass for infrastructure and research groups alike.

“Grafana Cloud was a no-brainer. We could do everything our on-prem setup did—and better—without having to manage it,”Jacob Lewis, Senior System Administrator, University of Iowa College of Public Health

A research institution with broad, critical infrastructure needs

When Senior System Administrator Jacob Lewis joined the team two years ago, he inherited a partially configured Grafana OSS deployment running inside the university data center. It collected everything, but visualized almost nothing. “It worked, but it had only ever been taken to about 75%,” he said. “Every single metric coming from some of our servers getting pulled in which 99% of it is useless.”

With increasing demands for reliability and the need for proactive insight across diverse systems, it was time for a change.

Too much data, too little visibility, and real risks during outages

The open source, on-prem instance had become a liability. It ingested hundreds of thousands of metrics, but provided no practical dashboards or alerting. Worse, the team realized that if the data center went down, they would lose all observability, including insight into their cloud resources.

A university-wide outage at a separate data center underscored the risk. “If our data center went down, we’d lose everything—alerting, dashboards, and visibility into AWS. Even though that’s its own problem, we shouldn’t lose observability on top of it,” Jacob said.

The team began evaluating alternatives, but the decision quickly became clear: They needed a cloud-hosted solution that reduced operational overhead and provided a future-proof path to unified observability across environments.

Migrating to Grafana Cloud and refining observability 

CPH moved from Grafana OSS—where they had been operating since February 2022— to Grafana Cloud,driven by both operational complexity and risk.

Their self-hosted Grafana environment had become increasingly difficult to maintain, requiring frequent restarts and creating a critical blind spot: if their infrastructure went down, they also lost visibility into the very telemetry needed to diagnose the issue. This dependency ultimately pushed the team to make an executive decision to transition to Grafana Cloud, prioritizing reliability, reduced operational burden, and consistent access to observability data—especially important for a distributed team frequently working remotely.

Following the move, CPH initially leveraged Grafana Cloud Metrics for metrics, while in parallel building out their Grafana Cloud Logs and Grafana Cloud Traces deployments for logs and traces, respectively.

Jacob initiated the migration with a “lift and shift” approach, pointing existing Prometheus scrapers to Grafana Cloud. Immediately, the team saw the volume problem clearly: more than 630,000 metric series, which would have cost more than $5,000 per month if left untouched.

To right-size their observability footprint, Jacob manually removed unnecessary metrics, focusing on only what delivered value. The result was dramatic: ingestion was reduced to ~4,000 active time series, well within Grafana Cloud’s free-forever tier and aligned to the college’s actual operational needs.

(Jacob jokingly described the process as “four painful weeks,” noting that it was only after the fact that he learned about Adaptive Telemetry, a Grafana Cloud feature that can do the same thing automatically.) 

With the signal-to-noise ratio fixed, the team immediately began creating dashboards, alerts, and predictive metrics views that were impossible on the old on-prem system. Grafana Cloud’s simplicity enabled rapid iteration. Jacob described how easy it is to add a metric and almost immediately start seeing reporting in their dashboard. 

“I instantly have data for anything I need to do almost immediately, which is a godsend,” Jacob said

Predictive insights, reliable alerting, and a single pane of glass

Migrating to Grafana Cloud transformed how CPH monitors its infrastructure. Key outcomes include:

A unified view of every system

Grafana Cloud now provides the single pane of glass CPH had been missing—aggregating on-prem, virtualized environments and AWS systems in one place.

“It really is life-changing,” Jacob said. “Seeing my entire environment and knowing that I’m getting correct statistics and knowing that everything’s accurate and up to date, it is eye opening.”

Proactive incident prevention

With reliable alerting and cleaner metrics, the team can better anticipate issues. Predictive rules help identify early warning signs, giving the three-person team more time to focus on higher-value work.

Reduced operational burden

By offloading backend maintenance, CPH no longer has to manage storage, upgrades, or high-availability concerns. Grafana Cloud simply works.

Cost-efficient observability

By eliminating excess ingestion, CPH reduced their projected observability costs by more than 98%—all without compromising visibility.

Expanding usage across research groups

CPH is now building dashboards for faculty and research teams, improving transparency and enabling data-driven decisions without IT acting as an intermediary. “We want to make things less of a black box,” Jacob said. “[Our end users] should be able to see it because it’s theirs and they should be able to get a comprehensive view of their equipment even though we run it.”

Looking ahead: Logs, cloud cost insights, and deeper AWS integration

The team’s next step is to bring Grafana Cloud Logs, powered by Loki, into the fold for centralized log management, continuing the journey to true centralized observability. They are also planning to build dashboards that ingest AWS usage and cost metrics directly into Grafana Cloud, reducing reliance on native cloud monitoring tools and enabling cross-platform correlation.

“We saw that it can do everything we need,” Jacob said. “I want to import every single usage statistic out of AWS and into Grafana. I don’t want to use AWS for monitoring—I want to use it for compute and use Grafana Cloud to correlate that telemetry with all my other systems”.

As the college continues to scale its research programs, Grafana Cloud provides the flexibility, reliability, and open ecosystem they need to grow without increasing operational burden.