Kubernetes Hosting

Self healing

Every component of the web stack has built in health checks, which automatically restart a service that has failed or is returning errors. This keeps applications online through heavy traffic periods, and recovers sites quickly in the case of component errors.

Ready to scale

Depending on deployment strategy built-in, virtually instant vertical and horizontal scaling is available to serve fluctuating traffic demands.

Redundant infrastructure

Fully redundant design means that applications quickly recover from component or hardware failure.

State based infrastructure as code

Every aspect of a service, from stack components to configuration files, is version controlled and state based. This means you know exactly what configuration is running at any given point and can revert to previous configuration to troubleshoot. Duplicating a service for development or staging provides like for like environments that behave identically.

CI / CD ready

M.D.G. IT Kubernetes hosting connects directly with pipeline deployments, allowing you to deploy code directly to persistent storage or using container deployment. Our support team provide free deployment pipeline support to get your application from version control to live at the click of a button.

Dedicated CPU & RAM

Unlike a dedicated or Virtual Private Server, 100% of service resources are available directly to the application, without kernel, file cache or operating system overhead.

Portable applications

Because the web stack is explicitly defined in code, applications are portable to other Kubernetes clusters without additional setup, preventing vendor lock-in to a specific hosting company.

Zero-downtime upgrades

During upgrades of stack components like PHP or Node.js, the system waits for health checks on the newer component to pass before tearing down the previous version, allowing true zero-downtime upgrades of the stack. Because each environment is replicable down to component versions and service configuration, upgrades can be tested completely in staging before being rolled out in production.

Advanced cron control

Cron jobs have their own isolated resource limits and advanced scheduling, allowing concurrency and maximum runtime to be configured per job. This means that the application doens't have to deal with cron race conditions, and crons do not compete with public facing services for resources.