From 9e4c96fe84e35db6ccbce14720bf829405fabcc1 Mon Sep 17 00:00:00 2001 From: zsamji Date: Mon, 18 Dec 2023 16:49:14 -0800 Subject: [PATCH] chore: updated from wiki removed api and added teams to yamlv2 --- wiki/Pathfinder-Uptime-Monitoring.md | 4 ---- 1 file changed, 4 deletions(-) diff --git a/wiki/Pathfinder-Uptime-Monitoring.md b/wiki/Pathfinder-Uptime-Monitoring.md index e0758ebd..0566af66 100644 --- a/wiki/Pathfinder-Uptime-Monitoring.md +++ b/wiki/Pathfinder-Uptime-Monitoring.md @@ -30,16 +30,12 @@ Going from our primary to fail over is something we are prepared for. You can ex ## Example Messaging 1. The Gold Keycloak _environment_ instance is in the process of failing over to the DR cluster - * The CSS APP is being put in Maintenance mode. Please check our [Uptime](https://uptime.com/statuspage/bcgov-sso-gold) before using our service. - 2. The Gold Keycloak _environment_ instance has failed over to the DR cluster - * DR deployment is complete, end users continue to login to your apps using the Pathfinder SSO Service (standard or custom). * Any changes made to a project's config using the Pathfinder SSO Service (standard or custom realm) while the app is in its failover state will be lost when the app is restored to the Primary cluster. (aka your config changes will be lost). * The priority of this service is to maximize availability to the end users and automation. 3. The Gold Keycloak _environment_ instance has been restored to the Primary cluster (aka back to normal). - * We are be back to normal operations of the Pathfinder SSO Service (standard and custom). * Changes made to a project's config using the Pathfinder SSO Service (standard or custom realm) during Disaster Recovery will be missing. * The priority of this service is to maximize availability to the end users and automation.