What is maintenance window notification policy?
Here is the 2026 maintenance notification policy for OCI:
For routine infrastructure maintenance (e.g., hardware upgrades or non-emergency firmware updates), OCI follows a 14-day lead-time policy:
14 Days Before: You receive the initial "Scheduled Maintenance" announcement via the OCI Console and email.
7 Days Before: A secondary reminder is sent if you haven't proactively rescheduled or taken action.
24 Hours Before: A final "Upcoming Maintenance" alert is triggered.
At Start/End: You receive real-time notifications when the maintenance window begins and successfully closes.
In 2026, OCI does not just rely on email. Maintenance alerts are integrated into the OCI Announcements and Notifications services:
Console Announcements: A high-visibility banner and a dedicated "Announcements" dashboard.
Email: Sent automatically to the Tenancy Administrator and any defined Infrastructure Contacts.
SMS & Slack: If configured via the OCI Notifications service, you can receive instant mobile or chat alerts.
PagerDuty/ServiceNow: Enterprise customers often pipe these alerts directly into their ITSM tools via HTTPS webhooks.
One of the most important aspects of the 2026 policy is your ability to "push back."
Reschedule Window: For most maintenance events, you can use the "Reschedule" button in the Console to move the update to a different time within a provided 14-day window.
"No Preference" vs. "Custom Schedule": For advanced services like Exadata, you can set a Maintenance Scheduling Policy that tells Oracle: "Only perform maintenance on the 3rd Sunday of the month at 2:00 AM UTC."
| Maintenance Type | Lead Time | User Action Allowed | Impact |
| Standard Routine | 14 Days | Yes (Reschedule) | Live Migration (Zero downtime) |
| Reboot Migration | 14 Days | Yes (Manual Reboot) | Brief Reboot |
| Security Patch | 7 Days | Limited | Depends on severity |
| Emergency Change | < 24 Hours | No | Potential brief interruption |
New in 2026, OCI uses AI Triage to tailor notifications. Instead of a generic "A server in Phoenix is going down," the alert is dynamically generated to say: "Your 'Blog-Server-01' in the Production Compartment is scheduled for a reboot. We recommend moving traffic to 'Blog-Server-02' before Friday."
To ensure these alerts don't go to an unmonitored inbox:
Navigate to Identity & Security > Identity > Tenancy Details.
Update the Maintenance Contacts list.
Pro-Tip: Add a distribution list (e.g., devops-alerts@yourcompany.com) rather than an individual person’s email to ensure the team is always informed.