If you mass subscribed then it would help to know what kind of hardware you're running Lemmy on, and what type of connection. Since DB and connectivity issues might explain the issue. Generally you can't "catch up" if your instance is unreachable for any length of time so being available 24/7/365 is important if you want to ensure you get and can see all content.

I am new to the fediverse, and could have misconfigured my docker image. But, I have de-sync issues the same way. What I started doing is browsing other Lemmy instances directly. Because my subscription does not seem to pull all content and all comments. It picks and chooses to sync about 4 or 5 post out of the ~50 a day from my 6ish different subscriptions.

I am not totally sure how the software talks back and forth between instances. But, it seems like a huge issue if instances don't have a feature to stay in sync.

You only receive content updates for the communities your local users are subscribed.

That being said federation has been struggling significantly since this past reddit exodus and there's time where outages of any kind might have made you lose federated content. I don't think is possible to "catch-up" all of what you have lost.

You can however per example search a port url of a remote instance to force it federate.

I use that periodically to compare feeds, and like I said sometimes a post or comment is missing, actually I often see a comment that looks like it’s responding to a another comment, but I cannot see the parent comment.

There have been bugs in Lemmy not sending comment deletes to all the instances. And lemmy.world and lemmy.ml were not communicating fro Saturday through Tuesday. Lemmy.world had some significant outages. It gets pretty tricky to track down and identify exact causes while things are unstable.

I ran into an issue similar to this while hosting my server. Turns out there is a ”feature“ (seems bugged) on my ASUS router called “DDOS Protection” that mistook inbound lemmy traffic as attacks and just dropped them instead of forwarding them. Once that was disabled everything worked like normal. Maybe be worth trying.

Be ready to use other methods to deal with DDOS though.