[MM-58355] Send invalidate cache message across the cluster so that websocket connections on other instances are invalidated correctly (#27204)

* [MM-58355] Send invalidate cache message across the cluster so that websocket connections on other instances are invalidated correctly

* Add suggestion to clear the session cache on the local node as well

* Force read from master DB when gettting channel members for websocket to avoid any DB sync issues

* PR feedback

* Missed generated files

server/channels/app/authorization.go

@@ -82,7 +82,7 @@ func (a *App) SessionHasPermissionToChannel(c request.CTX, session model.Session
 		return false
 	}
 
-	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(session.UserId, true, true)
+	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(c, session.UserId, true, true)
 	var channelRoles []string
 	if err == nil {
 		if roles, ok := ids[channelID]; ok {
@@ -134,7 +134,7 @@ func (a *App) SessionHasPermissionToChannels(c request.CTX, session model.Sessio
 		return true
 	}
 
-	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(session.UserId, true, true)
+	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(c, session.UserId, true, true)
 	var channelRoles []string
 	for _, channelID := range channelIDs {
 		if err == nil {
@@ -266,7 +266,7 @@ func (a *App) HasPermissionToChannel(c request.CTX, askingUserId string, channel
 	// We call GetAllChannelMembersForUser instead of just getting
 	// a single member from the DB, because it's cache backed
 	// and this is a very frequent call.
-	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(askingUserId, true, true)
+	ids, err := a.Srv().Store().Channel().GetAllChannelMembersForUser(c, askingUserId, true, true)
 	var channelRoles []string
 	if err == nil {
 		if roles, ok := ids[channelID]; ok {