diff --git a/docs/products/postgresql/concepts/upgrade-failover.md b/docs/products/postgresql/concepts/upgrade-failover.md
index ac14680f..e7210b82 100644
--- a/docs/products/postgresql/concepts/upgrade-failover.md
+++ b/docs/products/postgresql/concepts/upgrade-failover.md
@@ -33,8 +33,8 @@ different procedures in case of primary or replica nodes disconnections.
 ### Primary server disconnection
 
 If the **primary** server disappears, Aiven's management platform uses
-a **60-second timeout** before marking the server as down and promoting
-a replica server as new primary. During this 60-second timeout, the
+a **180-second timeout** before marking the server as down and promoting
+a replica server as new primary. During this 180-second timeout, the
 master is unavailable (`servicename-projectname.aivencloud.com` does not
 respond), and `replica-servicename-projectname.aivencloud.com` works
 fine (in read-only mode).
@@ -49,7 +49,7 @@ DNS is switched to point to the new replica server.
 ### Replica server disconnection
 
 If the **replica** server disappears, Aiven's management platform uses
-a **60-second timeout** before marking the server as down and creating a
+a **180-second timeout** before marking the server as down and creating a
 new replica server.
 
 :::note
@@ -103,7 +103,7 @@ procedure is followed:
 
 :::note
 - The old primary server is kept alive for a short period of time (minimum
-  60 seconds) with a TCP forwarding setup pointing to the new primary
+  180 seconds) with a TCP forwarding setup pointing to the new primary
   server allowing clients to connect before learning the new IP address.
 
 - If the service plan is changed from a business plan that has two nodes