forked from tsileo/microblog.pub
-
Notifications
You must be signed in to change notification settings - Fork 2
Open
Labels
bugSomething isn't workingSomething isn't working
Description
I woke up this morning, and my blog was really slow. Then I tried to SSH to it, but it was even slower. All that I could see was that a couple of these INV processes were eating the CPU:

The blog processes eventually died, and when I looked into the logs, it was covered in failed (timed-out) requests for federated user keys.
File "/mnt/volume_sfo3_01/microblog.pub/./app/httpsig.py", line 125, in _get_public_key
actor = await ap.fetch(key_id, disable_httpsig=True)
│ │ └ 'https://bcast.guru/actor#main-key'
│ └ <function fetch at 0x7fcae5127f40>
└ <module 'app.activitypub' from '/mnt/volume_sfo3_01/microblog.pub/./app/activitypub.py'>
##...
File "/usr/lib/python3.10/contextlib.py", line 153, in __exit__
self.gen.throw(typ, value, traceback)
│ │ │ │ │ └ <traceback object at 0x7fcae1b542c0>
│ │ │ │ └ ConnectTimeout(TimeoutError())
│ │ │ └ <class 'httpcore.ConnectTimeout'>
│ │ └ <method 'throw' of 'generator' objects>
│ └ <generator object map_httpcore_exceptions at 0x7fcae1ba6880>
└ <contextlib._GeneratorContextManager object at 0x7fcae1c982b0>
File "/root/.cache/pypoetry/virtualenvs/microblogpub-QDkw_Xeb-py3.10/lib/python3.10/site-packages/httpx/_transports/default.py", line 77, in map_httpcore_exceptions
raise mapped_exc(message) from exc
│ └ ''
└ <class 'httpx.ConnectTimeout'>
My working hypothesis is that it happened after I got a couple of new follow(er)s... I think this is related to some Federation protocol that is running havoc. That would not be the first time I heard of these cases. I will keep monitoring and investigating...
Metadata
Metadata
Assignees
Labels
bugSomething isn't workingSomething isn't working