adql

19 ms·

211.
by adql 3y ago
The "different set of services" is so basic it can be ran anywhere, and near-everywhere is cheaper than AWS
212.
by adql 3y ago
It sounds very much like what current CPUs do, with x86 being mapped into execution units dynamically via microcode, except (I assume ? article was very low on detail) in case of Crusoe it just did it for big VLIV instruction blobs.
213.
by adql 3y ago
A lot of people want to see it purely on hope that under MS Actiblizzard will be less shit of a company. Which is... optimistic.
214.
by adql 3y ago
One perfectly homogenous server gets 20MB POST with image that then needs to be reencoded/cropped Other gets tiny GET for CSS. Boom, now your equal servers are inequally loaded. The beauty of leastconn is that if 1st server is loaded i
215.
by adql 3y ago
Well, in short, yes. > I always thought that for hundreds of backends and above 10K RPS those algorithms scale poorly https://www.haproxy.com/blog/haproxy-forwards-over-2-million... > First of all, scaling beyond
216.
by adql 3y ago
Start here: https://github.com/haproxy/spoa-example https://www.haproxy.com/blog/extending-haproxy-with-the-stre... We used that to make an SSO login site that works independently on what is on th
217.
by adql 3y ago
It kinda fell out of favour when even midrange server can do tens of gigabits of TLS traffic. So you need to be very big traffic wise to make it worth.
218.
by adql 3y ago
...that's just push model. "Signalling" via "well, the loadbalancer have 10 sessions max per server" is enough. Pull model just adds unnecesary RTT. > Off the top of my head I'm trying to think how a host co
219.
by adql 3y ago
That's push model when per-server maxconn is full tho. The biggest benefit from pull model is not having to update backend server list every time you add/remove one but outside of that it isn't really all that beneficial. You
220.
by adql 3y ago
You need health checks. If not for LB then for your own monitoring.
221.
by adql 3y ago
"Wait, why this is dead for 4 weeks and not processing anything > One of the nicest things about pull load balancing is that it completely replaces the need for a separate registration + health check system. "Jeff, stop taking
222.
by adql 3y ago
> But there are some cases, such as when serving audio/video streams (looking at you, porn :) where the server will send its response straight to the user, not back through the load balancer. So, the load balancer doesn't know
223.
by adql 3y ago
And the fact that you should tune for loadbalancer to queue and eventually drop, not your app server. Then in case of leastconn there will be no dropped connections as long as total server capacity < incoming traffic.
224.
by adql 3y ago
It's not difficult. loadbalancer knows how many requests are in-flight to server . The OP is plainly wrong. That's why leastconn works in the first place. And leastconn is almost always one you want. It's almost magical. GC
225.
by adql 3y ago
> All push load balancing algorithms try to somehow predict how busy downstreams are. They KNOW how busy they are. They are the ones tracking and forwarding connections to them. That's why leastconn works in the first palce But, for
226.
by adql 3y ago
Leastconn also have very nice characteristics, that if your app say gets into long GC it will near-immediately stop sending connections there Also the "overrun" of dropped connections shouldn't really happen if * incoming con
227.
by adql 3y ago
Yeah and riding 1000cc bike is just like driving the bicycle with Extreme Speed /s > In the case of webRTC, I've seen so many projects do this where they’re like, “It‘s peer-to-peer! You just connect to the signaling server and
228.
by adql 3y ago
Your left hand doesn't turn into your right hand if you lie on the bed different way
229.
by adql 3y ago
...you just expect random readers to know how lung looks on (i assume) MRI, from (I assume) top ? I'd also assume it would be simple to spot just because the image is always took from the same direction so the narrow part points out t
230.
by adql 3y ago
That site is fucking garbage, and there doesn't even see to be a dock on what it supports. Like, programmable key is cool as an idea but I need smartcard support and a button on it to confirm transaction to replace YK usage...
231.
by adql 3y ago
This fucking article. CF, WHICH IS THE FUCKING SOURCE OF THIS PROBLEM, complains about the problem
232.
by adql 3y ago
the background makes it badly readable on phones too
233.
by adql 3y ago
Not talking about company being more successful but better to the actual customers. Google isn't exactly a good example here Also Microsoft went public almost 4 decades ago, apple went public over 4 decades ago, the landscape looked a
234.
by adql 3y ago
Those look very much as an exception to the rule
235.
by adql 3y ago
I don't think it's irrational, it rarely improves the service
236.
by adql 4y ago
> git reflog (git usage or git cmdlog) It kinda needs separate command that shows you both reflog entries and commands that lead to it. Maybe merged with undo command so you could look at the history of operations, go back to that point,
237.
by adql 4y ago
> Maybe it's just morning, or maybe I'm a luddite, or both, but my knee-jerk reaction to this was "why does anyone need this? Just learn git." Knowing the git enough to not need crutches is opposite of luddite. But ye
238.
by adql 4y ago
Git gud with git or get gut ?
239.
by adql 4y ago
Also FSD in 2024 will work perfectly
240.
by adql 4y ago
The 2020 was in negative, 2021 was also in negative. That's not "becoming profitable"

More ›