@j not trying to sound stupid but is it $156 USD a month to help support you guys?, dontcha have a $app for any amount of payments like we do? 😉 BOOSTED and APPRECIATE YOU
@EmilyBehave $156 is the total amount people are contributing every month so far. You can contribute any amount per month!
We can't use $app cause we're in Australia
@j like i said i wasn't trying to sound stupid ... i just was lol well I was not in the know with over seas apps or how much you want. But the links you posted help. I got it noted, and boosted!
You should post your financial goal to reach for support and your links 😉💳💵
Hugs from my heart to yours J and @e ❤
@EmilyBehave ❤️ ❤️
We're waiting for our costs to stabilise before we announce an amount as we've been around for less than two billing cycles, but it's about $800-$1400 AUD per month at the moment for just server costs alone. We're pretty much full time on Switter right now but it seems very unlikely we'd get enough in donations to pay ourselves :(
@j ❤ you just keep focusing on the good stuff and the good stuff will keep comin to you.
Thanks so much for sharing this info. 😚 thank you J - blessings and RICHES
@j Eugen said that mastodon.social costs him 40 euro per month in server costs. I'm surprised the cost is so high for Switter despite having fewer users.
Also, another thought. If your instance is primarily used for commercial activity, you can legitimately charge a fee for the service instead of depending on donations so that you can also support yourself working full time.
@j @metalbob 4x EX40-SSD. Nginx, 1 streaming process and 4x10 pumas on one; 200 sidekiq threads on another (each machine has its own pgBouncer); redis and postgres on another. I'm still in the process of setting the DB up, but the last server is gonna be a replica of the DB and I'm gonna split the redis into "persistent" and "volatile" to minimize disk i/o. There's also a completely separate, small machine with nginx proxying to wasabi object storage with proxy_cache_lock to reduce wasabi egress
@j Hmm. I have 150 threads for sidekiq in total. Haven't measured job throughput yet. Guess I'd have to hook into statsd to get that number..?
@Gargron I just eyeball the sidekiq admin panel for that one heh. I do want to have more detailed stats on count/duration stats for job types at some point too
@j Mastodon can do statsd metrics, and netdata supports being a statsd server, but I haven't tried it out yet. I remember a few months ago people had issues with the statsd code leaking sockets in the sidekiq process, I am not sure if that was fixed since or not but I am a bit cautious about enabling that
@j I'm not exactly sure, I know that the hardest part of statsd metrics is making them make sense in graphs
@Gargron No derivative function? The thing I can't seem to see netdata doing is being able to have dashboards and aggregate metrics from different machines. Like, I wanna see jobs per second by machine
@austintravels Sadly Patreon does not.. they recommend cancelling the pledge after the first month has gone through for once off payments which ain't great: https://patreon.zendesk.com/hc/en-us/articles/204606215-Can-I-make-a-one-time-payment-
@j No problem ! And if ever you guys start accepting gift cards i.e. amazon, etc please DO make that announcement !
@austintravels hmm interesting idea, sadly our hosting providers don't take gift cards but we'll think about it!
@j why do you need to add more servers?
maybe if someone wrote articles on why it costs, where costs go , breakdown of what you have to do, etc. fundraising efforts might be more effective?
if you give people details. Like why does it cost. like basic details then a CLICK for more in depth.
I know yall are stretched thin, but it might help.
sorry just my two cents. I am always working nonstop on my sites so I understand the pain
@KLttyKat occasional load spikes when popular accounts toot or boost which slows everything down. We haven't done so yet because the costs fluctuated a lot in the first two months hence the figure are all estimates at the moment
@j totally get that. figured that is why service was slow at times.
like a hundred people trying to get through a doorway.