[Rspamd-Users] Question about the current status of the public fuzzy service

[Vsevolod Stakhov]

Hello,

Please follow this ML posting policy and don't top post.

On 11/11/2025 16:17, Florian Piekert via Users wrote:
> Hello,
> 
> that is a question I ask myself as well. The "public" ones seem to hold 
> only 671 fuzzy hashes. (latest rspamd update from today installed)
> 
> sonne      rspamd.com       671
> local                    591706
> theater    rspamd.com       671
> local                    394058
> butterfly  rspamd.com       671
> local                    130530
> 
> Am 10.11.2025 um 22:30 schrieb Herbert Alexander Faleiros via Users:
>> Hi,
>>
>> On Mon, Nov 10, 2025 at 08:31:12PM +0000, Vsevolod Stakhov wrote:
>>> On 10/11/2025 19:45, Herbert Alexander Faleiros via Users wrote:
>>>> I would like to ask about the current status of the public fuzzy
>>>> service (fuzzy[12].rspamd.com).
>>>>
>>>> According to the Incident Disclosure - Public Service Temporary
>>>> Suspension report published on October 18 2025
>>>> (https://docs.rspamd.com/blog/2025/10/18/incident-disclosure), the
>>>> service was temporarily disabled due to hosting provider issues.
>>>>
>>>> However, even though the DNS entries no longer point to localhost (as
>>>> they did right after the incident), the servers still do not seem to
>>>> respond on port 11335, they accept packets, but there is no reply.
>>>>
>>>> Could you please confirm whether the public fuzzy service is still
>>>> suspended, or if it has been re-enabled in any limited form?
>>>>
>>>> Thank you for your time and for maintaining this project.
>>>
>>> It's up and running, if you get no reply, then you're banned for 
>>> overuse or
>>> are unlucky to have bad neighbours (e.g. ec2).
>>
>> thanks for the clarification.
>>
>> Given what you mentioned (either being banned for overuse or having
>> "bad neighbours"), is there any way to check which of these cases
>> applies to my IPs or prefixes?
>>
>> Thanks for your help,

There are more than 500k hashes in that storage so far. The statistics 
reported is likely broken because we use now a different update logic. I 
think it could be easily fixed one day.