Notices by SpaceLifeForm (spacelifeform@infosec.exchange)

@aral

It is a message to Putin and Xi.

@feditips

40 posts is too little, because I can get that in less than 10 minutes. And then the duplicates show up.

The problem is not simple. It is not just a matter of adjusting a knob. It will put more cpu load and memory pressure on server side to track duplicates.

It may be helpful to have a user setting that says do not show me boosts of a post where the timestamp of the original post is over 24 hours old.

Or 3 days, or 7 days. Let the user decide depending upon how often they read.

That is a usable knob that will not stress the instance, and save bandwidth.

Another idea: Have a button to flag that you saw it.

The software has no clue whether you actually read it or not.

The combination of the post age and the user flagging they read it will work wonders.

@feditips

Even if you do this, you will likely see redundant boosts that still clog your timeline.

@feditips

I would add hiding boosts as an option.

@mhoye

Desktop or mobile? #Debugging

@mhoye

Thanks. IOS seems to be a factor.

@feditips @MastodonEngineering

It would be helpful if they posted on #Mastodon more often. Only 5 posts with thousands of followers is not what you want to do if you hope to get some engagement and some input.

@feditips

Unfortunately it is not that simple.

The server could do it based upon age of when the user last saw the boost. The problem is that this adds more complexity to the server to manage. And, there is no guarantee that the user actually saw (eyeball read) the boost in the first place.

I think this is best managed client-side, and that requires the client to track which post was boosted, and then would likely not be over 24 hours before it is aged out of a cache. Though, if you have the ram, you could make the expiry time longer.

Needs to be client side, but continue to waste bandwidth as happens now.

It is important that the original post creation timestamp is NOT a factor.

@aral @ariadne @mastohost

The fails are high.

I do not recall seeing long delays of your posts previously.

@aral @ariadne @mastohost

Another thought I had but was too tired to post before I fell asleep:

The Sidekiq queue may be building up because of a resource issue on your instance. There may be a memory leak problem.

I would reboot, and then monitor the stats closely.

If the queue starts to clear better, for a while, I would suspect a memory leak.

@jerry probably has some hints.

@aral

Dropping the fake that you know of.

Is the one on d25.community also fake or just someone with the same name?