State of Rackspace Spot 2026
The State of Rackspace Spot 2026 is the inaugural annual report from Rackspace Spot. It analyzes how hundreds of organizations run production workloads on Rackspace Spot fully managed Kubernetes clusters and Virtual Machine Cloudspaces.
While Rackspace Spot offers both on-demand and spot instances, this analysis focuses on spot instances, which run on an auction-based spot market where users bid directly for compute capacity, with prices driven by real supply and demand. Spot pricing history shows that Amazon Web Services (AWS) once operated a similar auction market, before moving to opaque fixed pricing.
The data comes from tens of thousands of active workloads across hundreds of cloudspaces and the findings challenge the assumption that spot instances are only suited for interruptible or non-critical workloads.
The report shows that 51% of cloudspaces run stateful containerized applications and 96.8% of active users run entirely on spot with no on-demand fallback.
Databases, caching layers, and ML pipelines are running in production on spot compute resources, achieving real cost savings without sacrificing fault tolerance.
For engineering teams and FinOps leads evaluating spot instances for serious workloads, this report is a practical reference point. It covers workload architecture patterns, GitOps tooling adoption, resilience strategies, and operational maturity across hundreds of organizations running containerized applications on Rackspace Spot.
With a preemption rate of around 0.1% on the Rackspace Spot platform, teams no longer need to worry about interruptions. Spot instances are charged by the hour, making costs predictable and easy to track. For teams serious about reducing compute costs, the data suggests reserved instances are no longer the default answer.
For any team running containerized applications on Kubernetes and looking to cut cloud costs without trading away reliability, the State of Rackspace Spot 2026 is the data they've been waiting for.
Start running your production Kubernetes workloads on Rackspace Spot today.
Frequently Asked Questions
What types of Kubernetes workloads actually run on spot instances?
More than most teams expect. The assumption that spot instances are only suited to throwaway batch jobs is outdated. Deployments covering stateless microservices, APIs, and web applications make up the majority of workloads, followed by batch jobs and scheduled automation. More significantly, over half of cloudspaces on Rackspace Spot run stateful applications including MySQL, PostgreSQL, and Redis using StatefulSets. Organizations are not choosing between stateless and stateful on spot infrastructure. They are running both.
How mature is Kubernetes spot instance adoption in 2026?
More mature than most industry benchmarks suggest, with GitOps and stateful workload adoption running well above CNCF survey averages. On Rackspace Spot, adoption has reached a record high: 96.8% of users run their workloads entirely on spot instances.
How do organizations handle resilience and high availability on spot instances?
Teams run multiple replicas across zones for high availability, with the most critical services running four or more. They back this up with automated database backups, failover, and object storage as backup targets.
What technology stacks are teams running on Kubernetes spot instances?
Most teams running Kubernetes on spot instances use the same production-grade stacks they would on on-demand infrastructure. On Rackspace Spot, teams run Go-based microservices, MySQL or PostgreSQL for their databases, Redis for caching, and Prometheus and Grafana for observability.
Is it safe to run databases and stateful applications on spot instances?
Running databases on spot instances has traditionally been seen as risky due to the possibility of interruptions. On Rackspace Spot, 51% of cloudspaces run stateful applications in production, with teams using operator-managed databases that handle automated failover, backups, and replication. For teams who prefer not to manage databases on instances at all, Rackspace Spot also offers a managed DBaaS with PostgreSQL, automated backups, and high availability built in.
Are organizations using spot instances for Kubernetes worker nodes in production?
Yes, running spot instances as Kubernetes worker nodes is one of the most common ways teams cut compute costs in production. Because Kubernetes handles pod scheduling and rescheduling automatically, worker nodes are a natural fit for spot. On Rackspace Spot, 96.8% of active users run entirely on spot with no on-demand fallback. With a preemption rate of around 0.1%, teams are not worried about interruptions, making spot worker nodes a reliable and cost effective choice for production workloads.
