Maintenance of Computle Infrastructure
At Computle, we regularly patch the underlying hypervisors to ensure performance, security, and reliability. Patching takes place during our global maintenance window on Saturdays, with Sunday reserved for rollback if needed. The process involves placing virtual machines into a paused or hibernated state, allowing updates to be applied seamlessly without interrupting workloads.
Patches are applied across fault domains and are thoroughly tested to minimize risk. Any sign of failure immediately halts operations. Each site is patched independently, and while rare, Windows instances may reboot during the maintenance window.
Our global maintenance window is on Saturdays. If critical vulnerabilities are detected outside of this window, we will engage with the account leads and schedule an emergency maintenance window.
Global Maintenance Window and Rollback
Patching takes place during our global maintenance window, scheduled for Saturdays. If any issues are encountered during the patching process, Sunday is reserved for rolling back updates and restoring the environment to its previous state. This two-day window provides ample time to test and correct any potential issues while minimizing the impact on customer workloads.
Notification
During the maintenance window, we automatically alert the logged-in user. However, if they have disabled notifications,
Pausing and Hibernation
When virtual machines are placed into paused or hibernated states, all active processes and sessions are preserved, allowing us to update the underlying hypervisor infrastructure without disrupting workloads. This ensures that the machine resumes smoothly after the update. However, there is a rare chance that Windows instances may require a reboot during this process. While this occurrence is infrequent, it is essential for customers to prepare for the possibility of a reboot during the maintenance window.
Underlying Hypervisor Patching
The patching process focuses on both the virtualization technology and the host operating system. Updates target core elements of the hypervisor, including networking, storage, and resource management, ensuring optimal performance and the latest security enhancements. Each update is carefully tested within isolated fault domains to verify that stability and functionality are maintained throughout the process.
Fault Domain Isolation and Testing
To limit the impact of patching, we isolate the updates within fault domains. This strategy ensures that patches are applied to small sections of the environment at a time, allowing any issues to be identified before a wider rollout. Each fault domain is rigorously tested after patching to ensure there are no performance issues. Should any failure occur during testing, the entire process is halted until the issue is resolved.
Site-by-Site Execution
Patching is executed on a site-by-site basis, ensuring that only one location undergoes patching at a time. This approach reduces the risk of disruption across your global operations, as unaffected sites continue running smoothly while maintenance is performed. Each site is updated individually, ensuring that the patching process is localized and isolated.
DevOps-Driven Automation
All hypervisor patching is automated through our DevOps pipeline, ensuring consistent, efficient deployment of updates. The automation pipeline includes pre-patch validation, post-patch testing, and continuous monitoring for any failures. Should an issue arise, the system automatically pauses the patching process and, if necessary, triggers a rollback to ensure system stability.
Through this carefully controlled maintenance process, Computle ensures that your virtual environments remain secure, stable, and up-to-date with the latest performance and security improvements, while minimizing downtime and operational impact.
Last updated
