[Rspamd-Users] Messages learned as SPAM but still delivered as not SPAM

[Andrei Goldchleger]

Hi there,

I have been using Rspamd for several months and have been tweaking my 
settings, but no matter what I do I cannot go over 30% SPAM being 
delivered to my inbox (meaning that over 100 SPAM messages, Rspamd 
detects and flags 70).

I get a lot of similar SPAM messages - I have Rspamd set up to learn 
SPAM and HAM as I move messages in and out the SPAM folder. However, 
more often than not, when I flag a message as SPAM I see a log entry 
such as the following:

2022-01-10 18:30:43 #923631(controller) <93604f>; csession; 
rspamd_task_process: skip learning: 
<q514n4g4148643n48406j5c434a4s5d4s254m4 at wj.novaoportunidadesim.com> is 
skipped for bayes classifier: already in class spam; probability 100.00%

Which, if I understand correctly, means that the message is already 
considered SPAM by the classifier. However, the message is still 
delivered to the inbox without any BAYES* symbol.

I wonder if it is possible that the redis database got "poisoned" in any 
way, thus impacting Rspamd efficiency. Several months ago I had issues 
with storage capacity (saw a bunch of "OOM command not allowed when used 
memory > 'maxmemory'" in the logs), but I increased the redis database 
capacity and since then the messages disappeared.

Configuration follows below.

Thanks,

Andrei

-----

My current Rspamd version: 3.0

Configuration snippets (from rspamadm configdump):

actions {
     greylist = 5;
     add_header = 2;
     reject = 150;
}

classifier {
     bayes {
         expire = 2144448000;
         backend = "redis";
         cache {
             backend = "redis";
         }
         tokenizer {
             name = "osb";
         }
         statfile {
             spam = false;
             symbol = "BAYES_HAM";
         }
         statfile {
             spam = true;
             symbol = "BAYES_SPAM";
         }
         store_tokens = false;
         signatures = false;
         min_tokens = 11;
         min_learns = 200;
         learn_condition = "return require(\"lua_bayes_learn\").can_learn";
         new_schema = true;
         users_enabled = true;
         autolearn [
             -3,
             5,
         ]
     }
}