The Strapi Kubernetes Plugin is a Kubernetes integration for the Strapi headless CMS that lets you deploy and manage Strapi directly in your k8s setup. It brings the proven reliability and scale of Kubernetes to enterprise Strapi projects. Built by Notum, Strapi’s first globally listed Enterprise Partner.
Your Website Is Your Main Sales Channel
Your website is where customers decide whether to trust you and whether to buy. When it slows under a traffic spike or goes dark during a release, the cost lands the same day: lost conversions, not just goodwill.
Strapi gives you full control of the content platform. The enterprise question is how to keep it fast and online as traffic grows and releases ships, day in and day out. A headless CMS Kubernetes integration answers it by running Strapi the way the world's most demanding digital services run: as a resilient, self-managing fleet.
What It Takes To Run At Scale
Any system that protects the integrity of its data keeps a single source of truth for writes. That single trusted writer is a strength, not a constraint: it is what guarantees your content stays consistent and correct. The opportunity is to surround that one writer with many fast copies (replicas) that share the read load, so the single writer is never overwhelmed.
Running an application as a coordinated fleet of copies is exactly what Kubernetes was built to do. What Strapi needs is a way into that world, and that is what our plugin provides.
The Strapi Kubernetes Plugin: The Full Power of Kubernetes
The Strapi Kubernetes Plugin is a Kubernetes integration for the Strapi headless CMS. Strapi is designed around a single read-write database; the plugin extends it to connect read-only database replicas as well, so a single authoritative writer can be surrounded by as many read-only readers as needed. You then run that fleet on Kubernetes, which keeps it healthy and scaled the same way it already runs your other services.
Kubernetes already gives the world’s most demanding systems self-healing, elastic scaling, and zero-downtime operation. The plugin brings all of that proven capability to your Strapi enterprise project. It works with Strapi 5.40 and later on any database Strapi supports.
Kubernetes is the most capable way to run the fleet, but the read-replica connection, the one thing Strapi can't do on its own, lives at the database layer, not in the orchestration. So the same one-writer/many-readers setup also runs on other container orchestrators or on infrastructure you already operate. As Strapi's first globally listed Enterprise Partner, Notum built it to be production-ready from day one.
What Running On Kubernetes Unlocks
The capabilities below come from Kubernetes, not from the plugin itself. What the plugin does is make them reachable: it lets Strapi run as one authoritative writer alongside read-only replicas, which Strapi cannot do on its own. Standing up the Kubernetes layer that delivers these strengths is a matter of configuration that Notum can provide end-to-end so that you can take the entire running fleet, not just a plugin.
| Capability |
Why does it matter? |
| Self-Healing |
If a copy of the application crashes or the server under it fails, Kubernetes notices instantly and starts a fresh copy to replace it. The platform repairs itself, so an unexpected failure becomes a non-event instead of an outage, and no one has to be paged in the middle of the night. |
| Elastic Scaling |
Kubernetes automatically adds more copies when traffic surges and removes them when things quiet down. You always have the right amount of capacity, so a campaign spike or entering a new market becomes a scaling event rather than a re-architecture project. |
| Load Balancing |
Incoming traffic is evenly distributed across all running copies, so no single copy gets overwhelmed. Pages stay fast even under heavy load, and one busy copy cannot slow everyone else down. |
| Zero-Downtime Releases |
New versions roll out one copy at a time while the others keep serving, and if a new version misbehaves, Kubernetes automatically reverts to the last working one. You can ship updates and undo bad ones, without users ever noticing. |
| Zero-Downtime Maintenance |
When the underlying servers need maintenance, such as upgrades, reboots, or patching, Kubernetes first moves the affected copies elsewhere so the service keeps running. Necessary maintenance stays invisible to customers. |
Capabilities Beyond Web Resilience
Running on Kubernetes also offers advantages that extend beyond keeping the service healthy amid change. These are properties of Kubernetes itself.
| Capability |
Why does it matter? |
| Multi-Region Replicas |
Run one writing region with read-only regions alongside it. A CDN caches static content near users, but anything it has not cached, plus dynamic reads, still travels back to the origin; a regional read replica puts that origin close too, so even dynamic reads stay local. The plugin enables the regional read origin; the CDN, database replication, and routing complete the picture. |
| Declarative Management |
Instead of configuring servers by hand, you describe the setup you want once, and Kubernetes keeps reality matching that description. If anything drifts from the plan, it self-corrects, which means far less hands-on maintenance. |
| Vendor-Neutral, No Lock-In |
Kubernetes runs the same way on every major cloud and on your own servers. You are not tied to a single provider, so you can move or mix infrastructure without rebuilding the platform. |
What It Adds Up To
Put together, these capabilities deliver one clear business outcome: lower operational cost and risk. Failures recover on their own, releases and maintenance no longer threaten uptime, and capacity follows demand without a re-architecture project. Your teams spend less time keeping the lights on and more time delivering value on infrastructure you already trust and are not locked into, not even to Kubernetes, since the same setup runs on other orchestrators or the stack you already operate.
Proven in Production: Scenic Group
Notum built this architecture in production for Scenic Group, the Swiss luxury travel company behind Scenic Luxury Cruises & Tours and Emerald Cruises. The project migrated 12 regional Sitecore websites to Strapi. It consolidated them into two, one per brand, across six regions and five languages, with non-primary regions served by dedicated read-only instances while a single primary stayed authoritative. That kept the platform stable across every market without putting the core at risk, and it is exactly the model the Strapi Kubernetes Plugin now turns into a supported, repeatable product.
Why Notum
As Strapi’s first globally listed Enterprise Partner, we bring this Kubernetes-native model to enterprise Strapi projects as a supported, repeatable solution rather than a one-off engineering effort. If content is central to how your business earns and keeps customers, this is the foundation that keeps it always on: scaling under load, surviving failures, and shipping changes without downtime.
Give Your Website Enterprise-Grade Resilience
If keeping your content platform fast and online is on your roadmap, we are glad to help you think it through. Write to us at sales@notum.cz to learn more about the Strapi Kubernetes Plugin and to explore your options, whether that means Kubernetes, another setup, or the infrastructure you already run.
FAQ About the Strapi Kubernetes Headless CMS Plugin
Can Strapi run on Kubernetes?
Yes. Strapi runs on Kubernetes, and the Strapi Kubernetes Plugin is what makes it run well there. The plugin is an npm package you add to your Strapi project (Strapi 5.40 and later) that lets you run one writing instance alongside read-only replicas, so the CMS can scale and stay available like any other service on Kubernetes.
Can Strapi use read replicas or a read-only database?
Not on its own. Strapi is built around a single read-write database connection. The Strapi Kubernetes Plugin adds the ability to connect read-only database replicas, which is the core capability behind running one authoritative writer with a fleet of read-only readers.
What does the Strapi Kubernetes Plugin actually do?
It does one focused thing: it lets Strapi connect read-only database replicas, so you can run a single writer with a fleet of read-only readers. It does not configure Kubernetes itself. The operational strengths people associate with Kubernetes, such as self-healing and scaling, come from Kubernetes, which you or Notum set up to run that fleet. Because that work happens at the database layer rather than in the orchestration, the plugin is not tied to Kubernetes specifically.
Do I have to run it on Kubernetes?
No. The plugin works at the database layer, adding the read-only replica connection that Strapi lacks, so it does not depend on Kubernetes. Kubernetes is the most capable and most widely used way to run the resulting fleet, which is why this guide focuses on it, but the same one writer with many read-only readers can also run on other container orchestrators or on infrastructure you already operate. Keep in mind that the self-healing and autoscaling described above come from the orchestrator, not the plugin, so the operational strengths you get depend on what you run the fleet on.
How do you scale Strapi for high traffic?
You run Strapi as a fleet of read-only replicas behind a load balancer, and let Kubernetes add or remove copies as load changes. Read traffic is served by the replicas, so the single writer stays protected during spikes. The plugin makes the read-only replicas possible; Kubernetes handles the scaling once it is configured.
Can you run Strapi in multiple regions?
Yes. You can run a single writing region with read-only regions alongside it, for example writes in one region and read-only replicas elsewhere. The plugin enables the read-only regional instances. Distributing them, replicating the database across regions, and routing users to the right region are handled by your infrastructure.
What is the difference between a CDN and Strapi read replicas?
A CDN caches static content close to users, but anything it has not cached, plus dynamic requests, still goes back to a single origin. A read-only replica puts the origin itself close to each region, so even dynamic reads stay local. They work together: the CDN caches content, the replicas move the live origin closer.
Does the plugin work with any database, including PostgreSQL and MySQL?
Yes. It works with any database Strapi supports, including PostgreSQL and MySQL. The plugin adds the read-only replica connection on top of your existing database setup.
Which Strapi version does the plugin require?
Strapi 5.40 and later.
Does it lock me into a specific cloud provider?
No. It runs on any conformant Kubernetes, including Amazon EKS, Google GKE, and Azure AKS, as well as self-managed and on-prem clusters, and it is portable between providers. You are not tied to a single vendor, and you are not tied to Kubernetes either: because the plugin lives at the database layer, the same setup runs on other orchestrators or the infrastructure you already operate.
Do I need Kubernetes expertise to use it?
The plugin handles the Strapi side. The Kubernetes side is set up separately, and you can run that with your own team or have Notum deliver the whole thing, the plugin and the Kubernetes layer together.
Is the plugin self-healing and auto-scaling on its own?
Those behaviors come from Kubernetes once the fleet runs on it, not from the plugin by itself. Kubernetes replaces failed copies and scales the fleet with demand. The plugin is what lets Strapi run as that fleet in the first place.
What happens if the writing instance goes down?
The read-only replicas keep serving content, so visitors still see your site. Publishing pauses until the writer is back, and you can configure the writer for high availability on Kubernetes the same way you would any other critical service.
How do I get the plugin and how is it licensed?
The plugin is available commercially from Notum, and is planned for the Strapi marketplace as paid listings become available there. Notum can provide the plugin on its own or as part of a full delivery that includes the Kubernetes setup.