# I am the Watcher. I am your guide through this vast new twtiverse.
#
# Usage:
# https://watcher.sour.is/api/plain/users View list of users and latest twt date.
# https://watcher.sour.is/api/plain/twt View all twts.
# https://watcher.sour.is/api/plain/mentions?uri=:uri View all mentions for uri.
# https://watcher.sour.is/api/plain/conv/:hash View all twts for a conversation subject.
#
# Options:
# uri Filter to show a specific users twts.
# offset Start index for quey.
# limit Count of items to return (going back in time).
#
# twt range = 1 27
# self = https://watcher.sour.is/conv/376foda
@eldersnake Btw... You're the only Yarn.social pod that I'm aware of that's running on a Raspberry Pi. How is it all going? Resource consumption, performance, overall experience? Can we improve anything? 🤔
@eldersnake Btw... You're the only Yarn.social pod that I'm aware of that's running on a Raspberry Pi. How is it all going? Resource consumption, performance, overall experience? Can we improve anything? 🤔
@prologic Really good actually. See pic, resource usage barely registers. A clean initialized yarnd barely registers any memory usage at all (on my RPi that means hovering around 70MB RAM usage). With my feeds and everything loaded in, it jumps to about 120MB usage, but that's still nothing in the scheme of things.\n\n 
@prologic I noticed that I often get timeouts from eldersnake’s new pod. I hope I’m not overloading it or something. 😨
@prologic I noticed that I often get timeouts from eldersnake’s new pod. I hope I’m not overloading it or something. 😨
@prologic I noticed that I often get timeouts from eldersnake’s new pod. I hope I’m not overloading it or something. 😨
@eldersnake @prologic Whereas, had I have run Nextcloud on my RPi (PHP) it would be easily 500MB+ RAM LOL. That's why I run Nextcloud on a rented VPS. I'm pleased to say yarnd runs like a champ ;)
@movq (#376foda) It's okay, it's because my tunnel service has a limit of about 8 unique connections per 45 mins. Lol, it's not much I know. I wouldn't worry, I am working on it. It is _technically_ possible to host the Pagekite.py program on a VPS and just do it myself (therefore not beholden to Pagekite.net's limitations), but I haven't had any luck getting that to work yet for some reason.
@eldersnake Yeah that's nice to hear! Let me know though if anything is a miss or could be improved! I want the experience of running pods as seamless and painless as possible, or I've done a bad job 😂 #decentralised #social #media et all. 👌 CPU will very rarely get used until/unless you start uplaoding videos (_which requires ffmpeg to be installed btw_). Memory can be controlled by -I, --max-cache-items and -C, --max-cache-ttl btw. The defaults are pretty good. but feel free to tweak these lower if you have to or don't need the longer active cache
@eldersnake Yeah that's nice to hear! Let me know though if anything is a miss or could be improved! I want the experience of running pods as seamless and painless as possible, or I've done a bad job 😂 #decentralised #social #media et all. 👌 CPU will very rarely get used until/unless you start uplaoding videos (_which requires ffmpeg to be installed btw_). Memory can be controlled by -I, --max-cache-items and -C, --max-cache-ttl btw. The defaults are pretty good. but feel free to tweak these lower if you have to or don't need the longer active cache
@movq Probably we're all hitting his PageKit rate limits. Pods are _very_ likely to blame here, without some smarter fetching algorithms to back off when there hasn't been an update in a while, this will be problematic. I _should_ work on this. I'm getting to a point where I _think_ I can maybe come up with a smarter global feed cache with better per-feed fetch cycles that back off to a maximum and fetch more often for more active feeds. Just need time/bandwidth to do this 🤗
@movq Probably we're all hitting his PageKit rate limits. Pods are _very_ likely to blame here, without some smarter fetching algorithms to back off when there hasn't been an update in a while, this will be problematic. I _should_ work on this. I'm getting to a point where I _think_ I can maybe come up with a smarter global feed cache with better per-feed fetch cycles that back off to a maximum and fetch more often for more active feeds. Just need time/bandwidth to do this 🤗
@prologic Yeah all good! The only thing that was out of the ordinary I guess was that initial problem I had with the open registration flag not working, which I worked around. But if that's the first you've ever heard of it, maybe it was something odd in my config/build.
@prologic Yeah all good! The only thing that was out of the ordinary I guess was that initial problem I had with the open registration flag not working, which I worked around. But if that’s the first you’ve ever heard of it, maybe it was something odd in my config/build.
@prologic @movq Granted, most people don’t have such restrictive rate limits like mine (or at all) ATM. From what I’ve seen the network traffic isn’t too noisy, the 5 min syncing for example I assume is pretty conservative compared to a lot of services.
@eldersnake Hmm thanks for reminding me! Let me see if I can repro that now... Feel free to jump on LibraChat on #yarn.social if you wanna help me debug this in real time 👌
@eldersnake Hmm thanks for reminding me! Let me see if I can repro that now... Feel free to jump on LibraChat on #yarn.social if you wanna help me debug this in real time 👌
@eldersnake @movq
> the 5 min syncing for example I assume is pretty conservative compared to a lot of services
yeah well If-Last-Modified is used here, so I/O will be minimal depending on the extend of busy feeds you may or may not follow. All things considering this approach scales a lot better than #ActivityPub based stuff like #Mastodon -- Just read @jlj's experience with the #firehose he managed to get in to 🤣
@eldersnake @movq \n\n> the 5 min syncing for example I assume is pretty conservative compared to a lot of services\n\nyeah well If-Last-Modified is used here, so I/O will be minimal depending on the extend of busy feeds you may or may not follow. All things considering this approach scales a lot better than #ActivityPub based stuff like #Mastodon -- Just read @jlj's experience with the #firehose he managed to get in to 🤣
@eldersnake @movq
> the 5 min syncing for example I assume is pretty conservative compared to a lot of services
yeah well If-Last-Modified is used here, so I/O will be minimal depending on the extend of busy feeds you may or may not follow. All things considering this approach scales a lot better than #ActivityPub based stuff like #Mastodon -- Just read @jlj's experience with the #firehose he managed to get in to 🤣
@eldersnake Really having a hard time reproducing this... I need to probably talk to you a bit more about what happened 🤔
@eldersnake Really having a hard time reproducing this... I need to probably talk to you a bit more about what happened 🤔
@movq @prologic He switched to solar powering his website, might be related! 🤣 On a serious note, he has an inbound connection limit from https://pagekite.net/ as far as I understood.
@movq @prologic He switched to solar powering his website, might be related! 🤣 On a serious note, he has an inbound connection limit from https://pagekite.net/ as far as I understood.
@prologic gotcha, I'll be out a bit today but once I'm back in. Thinking I should try repro it myself, build a separate new yarnd build on the RPi and see what happens.