DevOps in Embedded? Not What You Think (But It's Changing)

#embedded #devops

When we talk about DevOps, most people immediately think of cloud-native development:

Microservices, APIs, containerized deployments, CI/CD pipelines, production monitoring, dashboards, logs, alerts — all owned by the same team that writes the code.

In that world, Dev and Ops are tightly coupled.

But what about embedded software?

We also hear a lot about CI/CD and DevOps in embedded development — but if you’ve ever worked in this field, you probably know: it’s a very different reality.

Let me walk you through how it works for many embedded teams today — and where it’s starting to shift.

The Embedded Dev Reality

In embedded, developers typically:

Work on firmware locally, testing on demo hardware at their desks.
Push their code to SCM.
Ensure the CI/CD pipelines go green.
Tag releases, write changelogs, and prepare documentation.

Then, separate teams step in:

A CI/CD team maintains the pipelines, manages connected hardware nodes, and keeps the infrastructure running.
A Release team ensures the firmware binaries get integrated into systems like SAP or Teamcenter, where factory tooling can pick them up and flash them into devices on the production line.

Once the device is in the customer’s hands?

The software is running, but there is no operations team like in cloud services.

No one is scaling services, monitoring uptime, or responding to incidents live.

So, can we really talk about DevOps in embedded?

For many teams: Not really.

What we often call “DevOps” in embedded is more about:

CI/CD automation.
Infrastructure maintenance.
Pipeline engineering.
Managing test devices or Kubernetes runners.

But Here's the Twist: Some Embedded Teams Do Real DevOps

Recently, I learned that some embedded teams have crossed the boundary into true DevOps territory.

How?

These teams:

Use Over-the-Air (OTA) updates to continuously improve devices in the field.
Monitor log data streamed from connected devices.
Build dashboards to understand behavior, detect bugs, and track performance.
Instrument new logging points in firmware when they need better diagnostics.
Own both development and operations — reacting to what’s happening in the real world.

This is real DevOps:

Build it. Ship it. Monitor it. Fix it. Improve it. Repeat.

And it's exciting to see embedded development adopt these modern practices — even if the infrastructure and tooling look quite different from the cloud world.