Update about slow game post federating

Recently, when I post the new #HashTagGames post at 9PM est, it has taken 15 minutes or longer for all instances on the #Fediverse to get it. I determined the reason why is because my #ActivityPub #WordPress enabled site is getting overwhelmed  due to the sheer number of instances trying to communicate with us when the game goes out.

I have taken some measures to prevent this and tonight we will see how it goes. If it is still slow, I will continue to tweak things.

Just know, the game is going out on time, the post should arrive shortly each evening, and I will try to get this resolved one way or another.

If you have any tips on ActivityPub and WordPress, let me know at my personal account below.

Your host, @paul

@paul@OldFriends.Live How to play #HashTagGames Write something awesome, Use the Hashtag, Toot/Post and Repeat! Please Boost Hashtag Games on Mastodon and the entire Fediverse. Alert: Message to game players Every Night, 9PM EST, (6PM PT / 2AM GMT / 3AM CET / 1PM AEDT / 3PM NZST) Proudly hosting daily games since November 16, 2022" width="600" height="336" />

@hashtaggames

"Recently, when I post the new #HashTagGames post at 9PM est, it has taken 15 minutes or longer for all instances on the #Fediverse to get it ... why is because my #ActivityPub #WordPress enabled site is getting #DDoS due to the sheer number of instances trying to communicate with us when the game goes out."

I often check for details when I see posts like this, in the off-hand chance I might learn something

You seem to be on a single-user host via Namecheap

Assuming you're posting using TCP/IP, each packet you send out to the Fediverse requires an ACK packet in return before it continues, and most transactions involve scores of packets

Have you used something like Wireshark to look at exactly what you're sending out and receiving?

I doubt you're being DDoSsed so much as your very small instance is suddenly talking to tens of thousands of other instances, each one of which is trying to talk back to you all at once

Shorter: the Internet is not instantaneous

cc @paul

@hashtaggames

"Recently, when I post the new #HashTagGames post at 9PM est, it has taken 15 minutes or longer for all instances on the #Fediverse to get it ... why is because my #ActivityPub #WordPress enabled site is getting #DDoS due to the sheer number of instances trying to communicate with us when the game goes out."

I often check for details when I see posts like this, in the off-hand chance I might learn something

You seem to be on a single-user host via Namecheap

Assuming you're posting using TCP/IP, each packet you send out to the Fediverse requires an ACK packet in return before it continues, and most transactions involve scores of packets

Have you used something like Wireshark to look at exactly what you're sending out and receiving?

I doubt you're being DDoSsed so much as your very small instance is suddenly talking to tens of thousands of other instances, each one of which is trying to talk back to you all at once

Shorter: the Internet is not instantaneous

cc @paul

@FinchHaven Thanks. I changed the wording to the post. I just started troubleshooting the issue this morning and did a few things until I can break it down later overnight and troubleshoot more.

I added caching and going to add persistent object caching soon. I also notice php had low memory permissions so it was bottoming out quickly. I fixed that.

The WordPress instance sat idle until mid-February which explains why the jump is Latency.

I'm hosting on a VPS server through Namecheap at this time. Ubuntu 22.04, 12 GB ram, 8 CPU cores.

@HashTagGames

@paul

Good luck with it

Those were just my first, easiest thoughts

Dropped out of network troubleshooting a good long while ago...

cc @hashtaggames

@FinchHaven I am second guessing my choice of WordPress. I have been reading where people say one post or share from someone with a lot of followers will bring down your site each time... I am hoping, although some of the reading was recent, that devs have fixed it since then.

I am confident that the issues I am having is more towards my current, default WordPress configuration and not network related or even WordPress ActivityPub's fault.