Portal:Cloud VPS/Admin/Maintenance notification
When scheduled system maintenance will cause service outages, we try to provide our users with advance notice of the downtime. For serious outages, we also try to provide advance notice to our users' users. This table is meant to provide general guidelines but does not represent a formal commitment between WMCS admins and users. Our uptime stats may be good, but our uptime commitment is 'best effort'.
All maintenance tasks should be logged in the server admin log. Project-specific tasks are to be logged per-prioject, Toolforge actions should be logged on the tools log, and cloud-wide actions should be logged in the admin log.
Most advance notices are provided to users via mailing lists: cloud-announce, wikitech-l, and wikitech-ambassadors.
| Scheduled maintenance,
no expected downtime |
< 5 minutes of downtime,
restarts or resets |
expected downtime >= 5 minutes | expected downtime > 1 hour | |
|---|---|---|---|---|
| Select user-managed tools, services, projects | SAL | SAL | 1 day in advance:
|
7 days in advance:
|
| WMCS-managed project
(e.g., PAWS) |
SAL | SAL | SAL | 1 day in advance:
|
| OpenStack control plane | SAL | SAL | SAL | 1 day in advance:
|
| All of Toolforge,
Toolforge grid, or Toolforge k8s |
SAL | 1 day in advance:
|
7 days in advance:
|
7 days in advance:
|
| All of Cloud VPS
(including Toolforge) |
SAL | 1 day in advance:
|
7 days in advance:
|
7 days in advance:
|