Overview
The VPS platform is designed to minimize disruption during maintenance and to recover automatically from common infrastructure failures. While the platform provides a high level of resilience, it’s important to understand what is handled automatically and what remains the customer’s responsibility.
This article explains:
-
How planned maintenance is handled
-
What high availability (HA) means in practice
-
How host failures are recovered
-
The limits of platform-level protection
Planned Maintenance
From time to time, maintenance is required to ensure platform stability, security, and performance.
During planned maintenance:
-
Maintenance windows are announced in advance
-
Live migration is used whenever possible
-
Instances are moved between hosts to avoid downtime
???? Advanced
Live migration depends on a healthy guest OS and required guest tools. Instances without proper configuration may experience a brief reboot instead of a live migration.
High Availability (HA)
High availability is designed to protect against infrastructure-level failures, such as physical host issues.
HA provides:
-
Automatic detection of host failures
-
Automatic restart or migration of affected instances
-
No manual customer intervention required
⚠️ Important
HA protects against host failure, not against:
Application crashes
OS misconfiguration
Data corruption inside the instance
Live Migration
Live migration allows an instance to move between physical hosts with minimal or no downtime.
Live migration requires:
-
A supported operating system
-
Functional
qemu-guest-agent -
Proper OS configuration
???? Advanced
Instances missing required guest tools may be restarted instead of migrated to preserve platform stability.
Host Failure Recovery
If a physical host fails unexpectedly:
-
The platform automatically detects the failure
-
Affected instances are restarted or recovered on another host
-
Recovery occurs without customer action
⚠️ Note
While recovery is automatic, a brief service interruption may occur depending on the failure scenario.
What HA Does Not Do
High availability does not provide:
-
Application-level redundancy
-
Guaranteed zero downtime
-
Protection from user error
-
Protection from malware or ransomware
???? Advanced
Applications requiring zero downtime should be architected with redundancy at the application layer.
Customer Responsibilities
To fully benefit from HA features, customers should:
-
Use supported operating systems
-
Install required guest tools
-
Maintain OS health
-
Implement proper backup strategies
⚠️ Reminder
Platform resilience complements, but does not replace, good system administration practices.
Summary
-
Planned maintenance is announced and uses live migration when possible
-
HA protects against host-level failures
-
Host recovery is automatic
-
Guest OS health and configuration affect HA behavior
-
Application-level resilience remains the customer’s responsibility
Related Articles
-
Snapshots & Backups
-
Networking & IP Management
-
Resource Changes & Lifecycle Limitations
-
Support Scope & Responsibilities
