You can try to find a possible cause by enabling verbose logging:

SOGoDebugRequests = YES;
SOGoEASDebugEnabled = YES;
ImapDebugEnabled = YES;

LDAPDebugEnabled = YES;
or
MySQL4DebugEnabled = YES;
or
PGDebugEnabled = YES;

Check sogo.log for the first occurrence of "NSNull(misc) forwardInvocation:" after starting sogod with above parameters.
You say that the message is reported also for EAS-Ping-Requests … all requests or just for a specific user?

Well : I can't for now restart sogo on prod server, but I activate debug on a test server.
The error seems always linked to this mysql query, when we refresh the UI (= imap fetch) :
2017-11-02 09:45:09.635 sogod[29521] <MySQL4Channel[0x0x560c39ef83f0] connection=0x0x560c3a080d68> SQL: SELECT * FROM sogo_users_domainfr WHERE (c_uid = 'myuser') OR (mail = 'myuser') OR (mail1 = 'myuser') OR (mail2 = 'myuser') OR (mail3 = 'myuser') OR (mail4 = 'myuser') OR (mail5 = 'myuser');
2017-11-02 09:45:09.637 sogod[29521] <MySQL4Channel[0x0x560c39ef83f0] connection=0x0x560c3a080d68> query has results, entering fetch-mode.
2017-11-02 09:45:09.637 sogod[29521] ERROR(-[NSNull(misc) forwardInvocation:]): called selector stringByTrimmingSpaces on NSNull !
2017-11-02 09:45:09.637 sogod[29521] ERROR(-[NSNull(misc) forwardInvocation:]): called selector stringByTrimmingSpaces on NSNull !

which result is:
| c_uid | c_name | c_password | c_cn | mail | mail1 | mail2 | mail3 | mail4 | mail5 | domain |
+--------+------------------+------------------------------------+--------+------------------+----------------+-------------------+----------------+-------+-------+-----------+
| myuser | myuser@domain.fr | $1$oeytwCCw$3qjtf7GPCNRgM6gEt5Yv6/ | myuser | myuser@domain.fr | myuser1@domain.fr | myuser2@domain.fr | myuser3@domain.fr | NULL | NULL | domain.fr |

the "NULL" results are mail4 and mail5 (the user as only 3 aliases), I guess this is the problem : here the ERROR appears 2 times, when on the prod server, the users have no aliases, and it appears 5 times. I suppose you must test the field before pass it to stringByTrimmingSpaces ?

Unfortunatly, I couldn't reproduce the problem with EAS. Connecting a device with activesync on the test server does not produce errors as on the prod server. To be confirmed later if the error is fixed for the imap fetch.

Please try the latest nightly build and let me know if it solves the issue.

Hi,
as I want to only install builds via rpm to keep a compatibility between test server and prod server, can you confirm that I can just apply manually the modifications in this two files to test it, or not ?

And can you tell us when the version 3.2.11 will be available in the repos ?
thank you

You can install the latest nightly RPM builds.

See https://sogo.nu/download.html#helper

Oh sorry, I forgot to make a difference between "last version" and "last nightly build" :) so I was waiting for 3.2.11.
OK I test it.

Well, it seems good ! I continue to check the log for some days.

Just, the log being more clear now, I noticed a message "parseTimeZone: cannot parse time notation 'GMT'", I think that it is not in relation with this last built but since last update (3.2.9 -> 3.2.10) but had not seen it before (message hidden into all this stringByTrimmingSpaces...) May I open a new ticket, as I didn't found reference of this error in bug reports since v1.3.xx ? Or would you have a (quick) solution ?

If you isolate the event that causes the error, than please open an ticket and attached the event source (vCalendar).