Turning telemetry into action

Turning telemetry into action

Like many IT organizations, we collect a lot of telemetry.

My team is responsible for NetApp Active IQ. We take telemetry from more than 300,000 systems across the NetApp installed base and turn that raw data into insights that our customers, partners, support teams, and internal teams can use.

That amounts to roughly 10 terabytes of new telemetry every month, contributing to a data lake of approximately 4 petabytes across 24,000 customers.

At that scale, collecting data simply because we can isn’t a strategy. It has to have value.

In fact, a CFO has asked me directly: What are we getting from all this data?

It’s a fair question, and one that every organization collecting telemetry should be able to answer.

From reactive support to proactive insights

What we do today has its roots in AutoSupport. One of our earliest use cases was straightforward: detect a hardware problem and respond. If a disk or fan failed, telemetry could trigger the support process and help get a replacement part to the customer.

Since then, we’ve expanded how we use that data.

Today, telemetry helps our support teams troubleshoot problems and identify potential risks before they become larger issues. We also use it to help customers strengthen security, optimize capacity, plan their environments, and get more from their NetApp investments.

Over time, we’ve progressed from using telemetry reactively to using it proactively and predictively. Now AI is helping us take the next step.

Collect what you need—not everything you can

More data doesn’t automatically mean more value.

Telemetry costs money to collect, move, process, and store. Even though we’re a storage company, my team doesn’t get storage for free.

That’s why we continually evaluate what we collect.

We work with our engineering and product teams as they develop new releases and capabilities to determine what additional signals we’ll need. At the same time, we look at existing data to determine whether it still serves a purpose.

We actually have telemetry on our telemetry. We can see whether data is being accessed, which tools access it, and which APIs call it. If something is sitting there unused, we can determine whether we still need to collect and retain it.

The objective isn’t to build the biggest possible data lake. It’s to make sure the data we keep has a purpose.

Making the value visible

Collecting useful data is only part of the equation. We also want customers to understand what that data is doing for them.

That’s one of the ideas behind Value Insights in NetApp Digital Advisor.

We can use telemetry to show customers the value they’re receiving from capabilities such as storage efficiency, identify security risks that have already been mitigated, and highlight areas where additional action could improve their environment.

Capacity provides another example. We can identify cold data sitting on higher-cost all-flash storage and recommend options such as FabricPool to move appropriate data to a lower-cost tier.

The important part is connecting an insight to an outcome. It’s not enough to tell someone what’s happening in their environment. We want to help them understand what they can do next.

Moving from recommendation to action

This is where AI becomes especially interesting.

We’ve been applying AI and machine learning to telemetry for years to make predictions and identify patterns. Generative and agentic AI allows us to go further.

Instead of simply detecting a potential issue, we can determine the next best action. And instead of stopping at a recommendation, an AI agent could eventually take that action with the appropriate supervision and guardrails.

That’s where I believe AIOps is heading: shortening the distance between identifying a problem and solving it.

But we’re approaching that evolution carefully. We start internally, learn from our own users, and ensure the right guardrails are in place. From there, we work with trusted early-access users before expanding through a ring-based deployment model.

Start with lower-risk use cases. Learn. Get feedback. Then expand.

At INSIGHT, we’re introducing two ways to put Active IQ intelligence directly into users’ workflows. Digital Coach delivers AI-powered guidance within Active IQ, while Active IQ MCP brings NetApp intelligence to the AI assistants and agents users already choose. Together, they make it easier to turn insights into action wherever people work.

Start with the problem you’re trying to solve

For organizations looking to get more value from telemetry, I recommend starting with the use case.

What problem are you trying to solve? What data do you need to solve it? What outcome will tell you the investment is delivering value?

Support is often a natural starting point. From there, telemetry can help optimize the user experience, improve security and capacity management, and uncover opportunities that would otherwise be difficult to see.

AI can accelerate that progression, but it still depends on having the right foundation. You need the right data, secure access, and appropriate guardrails around how it’s used.

The opportunity is to turn the data you already have into something useful.

First, an insight. Then a recommendation. And increasingly, an action.

Listen to the podcast or explore the full podcast series.