If I didn't have AWS credits, it'd be ~$500. Keep in mind, I haven't really optimized for expenses and if I do, it'd probably be more in the territory of $300.
SN Item #1: Jun 11, 2021
SN Item #5,000: Nov 15, 2021 (151 days since Jun 11, 2021) 33 / day
SN Item #10,000: Feb 3, 2022 (80 days since Nov 15, 2021) 62 / day
SN Item #15,000: Mar 17, 2022 (42 days since Feb 3, 2022) 119 / day
SN Item #20,000: Apri 14, 2022 (28 days since Mar 17, 2022) 178 / day
SN Item #25,000: May 3, 2022 ( 19 days since Apr 14, 2022) 263 / day
SN Item #30,000: May 20, 2022 (17 days since Mar 3, 2022) 294 / day
It still needs some work but its functional. Speaking of which, anyone know a good alternative to connecting to my personal node over Tor? I'm thinking about getting a larger VPS and deploying both a dedicated lightning node and the site in docker containers.
I also plan to make this even more private by making a Tor version of the site and replacing the links to Twitter with links to a Nitter instance.
In the SN repo on Github, when an Issue is "closed as completed" with the message "Fixed locally", what does that mean? The issue persists yet. Are fixes made and then only pushed to production at a later time or something?
I'd be interested in how "busy" the SN backend is? If you look at stdout or something are the logs just racing through almost unreadable? Or like every few seconds? e.g. how often is the backend sending out a login e-mail? Because I do every few days.
I hope this isn't a weird question, I'm just interested in the reality of tech at this scale. Are you looking things like this up with very specific SQLs (because I'd for sure do that all the time)?
We have about 150 requests per second hitting the servers last I checked as reported by the load balancer.
I have slow logs to try to surface slow things and a bunch of dashboards about server memory, CPU usage, request latency, network download/upload, etc.
I often have multiple tabs open, each of those making 1QPS (even when not in focus), so it adds up.. Since the page has around 15 active users at a time, I think multiple folks do the same with multiple tabs open.
Is the higher QPS causing any notable issues or is it fine for now?
It appears to be fine. It occupies the db a bit, given some of the queries are expensive (eg looking for replies of replies), but I see that more of a query optimization problem than a request/second problem.
Force closing channels still happen and I really don't know how to avoid them.
I feel so powerless that can't control from my node this aspect.
I fucking hate these force closings. Are totally useless and idiotic.
Something must be changed in LN code, to have strict rules for different situations.
Why when there is more than 1 pending HTLC/channel, the channel seems to be "offline"? Reconnecting to peer seems to be a quick fix, but sometimes is not and channel goes to force close.