You can close, but I keep the patch for my case. Using Outlook 2016/365, when I delete all mails from Trash (> 3000 mails), I lost the sync and I need to reset the client completly (delete \AppData\Local\Microsoft\Outlook\myaccount.ost and AppData\Local\Microsoft\Outlook\EasCache\clientid.bin).
I don't have anymore error in my log. What I do, in /etc/z-push/z-push.conf.php define('SYNC_MAX_ITEMS', 5); No more problem at all, but I keep my patch because it avoid to get completely out of sync. zimbraHttpDosFilterMaxRequestsPerSec is not a solution, because it will impact all requests not only from z-push.
zimbraHttpThrottleSafeIPs: 178.xxx.xxx.xxx/32 zimbraHttpThrottleSafeIPs: 10.xxx.xxx.xxx/16 zimbraReverseProxyUpstreamReadTimeout: 3600s httpclient_soaphttptransport_so_timeout = 3600000 zimbra_session_limit_soap = 20 zimbra_session_timeout_soap = 3600 zimbra_waitset_max_per_account = 20 zimbraHttpDosFilterDelayMillis: -1 zimbraHttpDosFilterMaxRequestsPerSec: 30 Yes, I did many fixstates and clearloop during my debug. For information, I get that with only one account and one client, so I don't imagine...
Sorry remove "**".
WORKAROUND SOAP error