<![CDATA[#ActivityPub question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)?]]> question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)?

]]>https://bb.devnull.land/topic/f3d07add-1eba-4f11-a39a-ad986cb763f3/activitypub-question-is-it-ok-to-run-pruning-of-dead-inboxes-target-nodes-for-delivery-after-grace-period-of-e.g.-7-days-when-transport-failures-occur-repeatedly-timeout-ssl-dns-redirect-loop-unresolvable-host-400-404-410-all-5xxRSS for NodeMon, 24 Aug 2026 23:00:24 GMTMon, 24 Aug 2026 14:16:35 GMT60<![CDATA[Reply to #ActivityPub question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)? on Mon, 24 Aug 2026 17:49:06 GMT]]>@silverpill I think I'll make it configurable.

]]>https://bb.devnull.land/post/https://fosstodon.org/users/dusoft/statuses/117151743792484294https://bb.devnull.land/post/https://fosstodon.org/users/dusoft/statuses/117151743792484294Mon, 24 Aug 2026 17:49:06 GMT<![CDATA[Reply to #ActivityPub question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)? on Mon, 24 Aug 2026 17:42:56 GMT]]>@dusoft It is OK to prune but I think the period of 7 days is too short. The server admin might be on a vacation.

]]>https://bb.devnull.land/post/https://mitra.social/objects/01a034de-4112-7a90-91db-64e2b119545ahttps://bb.devnull.land/post/https://mitra.social/objects/01a034de-4112-7a90-91db-64e2b119545aMon, 24 Aug 2026 17:42:56 GMT<![CDATA[Reply to #ActivityPub question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)? on Mon, 24 Aug 2026 14:21:16 GMT]]>@julian That's helpful, thanks. I have implemented pruning, because all requests in the end have some costs. Of course, I could handle it also in softer way, e.g. 7 day grace period, then no-delivery list, then retry after 30 days... Then further incremental backoff.

]]>https://bb.devnull.land/post/https://fosstodon.org/users/dusoft/statuses/117150926537329268https://bb.devnull.land/post/https://fosstodon.org/users/dusoft/statuses/117150926537329268Mon, 24 Aug 2026 14:21:16 GMT<![CDATA[Reply to #ActivityPub question - is it OK to run pruning of dead inboxes (target nodes for delivery) after grace period of e.g. 7 days when transport failures occur repeatedly (timeout, SSL, DNS, redirect loop, unresolvable host), 400, 404, 410, all 5xx)? on Mon, 24 Aug 2026 14:18:27 GMT]]>@dusoft@fosstodon.org I don't think there is consensus on this particular point. Each implementation does their own handling.

e.g. Lemmy will queue messages for delivery forever (probably), but will place your instance on a "do not deliver" list for some time if delivery failures reach a certain threshold.

NodeBB's is more naive, it will try delivering per an exponential backoff forever, too.

]]>https://bb.devnull.land/post/https://activitypub.space/post/2384https://bb.devnull.land/post/https://activitypub.space/post/2384Mon, 24 Aug 2026 14:18:27 GMT