mrkurt

69 ms·

841.
by mrkurt 7y ago
That’s unusual, will you email me or support at fly.io so I can have a look? We use that KeyCDN test pretty frequently with different results. Those TLS handshake times aren’t great, I think that was probably the first load from Vault on ce
842.
by mrkurt 7y ago
This is a reasonable take. Let’s Encrypt is amazing and we don’t want to diminish their importance at all. We charge for certificates because the infrastructure to make SSL work (even when the certificates themselves are free) is complicate
843.
by mrkurt 7y ago
Ahhhh thank you! That's an incredibly nice thing to say.
844.
by mrkurt 7y ago
Stackpath is cheaper on bandwidth, mostly because of scale. For video based applications, this is a big deal (and we work with video companies to try and get better bandwidth pricing). For typical web app servers, bandwidth is usually not a
845.
by mrkurt 7y ago
That's actually a good, generic way to speed up the initial app connections. You can even do it with a CDN in front of a backend on a different network, since most CDNs pool connections. The problem is that everything useful a server s
846.
by mrkurt 7y ago
Ads would be terrible but I'm really interested in the whole idea of "do something else when an app os over budget". I hadn't considered much except a 402 status code, which seems kinda mean: https://developer
847.
by mrkurt 7y ago
That’s a pretty good definition of FaaS.
848.
by mrkurt 7y ago
Running your own bgp in multiple data centers requires some reasonable network engineering skills. Anycast, specifically, is a complex beast. If you use multiple transit providers, you have to continuously tweak things to make sure people a
849.
by mrkurt 7y ago
That sounds like the definition of “utopia” to me.
850.
by mrkurt 7y ago
How are you liking gVisor? We started with Firecracker and only spent a little time with gVisor, but it seems really nice.
851.
by mrkurt 7y ago
Packet charges for outbound bandwidth. The places you can get close to free on outbound bandwidth don’t give you the ability to do anycast and tend to over subscribe their networks. We’d like to grt network prices down but we can’t run our
852.
by mrkurt 7y ago
Algolia is great, we use it for search.
853.
by mrkurt 7y ago
This is my favorite Hacker News comment of all time.
854.
by mrkurt 7y ago
We got hung up with the migration tooling for popular frameworks. If we can get those migrations to work with minimal drama, we want to basically show people “global full stack” with app + cache + database.
855.
by mrkurt 7y ago
Only a little. It works very well, once we port our edge js library ( https://github.com/superfly/edge ) to Deno we will move over some things handling hundreds of millions of requests per day.
856.
by mrkurt 7y ago
FWIW, I love CloudRun and when I went to try and deploy Docker images in different places to see how we stacked up, it was the best of the big cloud offerings. I'm not sure any AWS PMs have ever even used Fargate ...
857.
by mrkurt 7y ago
Consul runs over a different connections with mutual TLS auth. That's the project we use!
858.
by mrkurt 7y ago
We’re moving the JS apps to just run on Deno containers. Deno is fabulous.
859.
by mrkurt 7y ago
Doh, missed that. We're not doing eBPF it's just user land TCP proxying right now. This will likely change, right now it's fast enough but as we get bigger I think we'll have more time to really tighten up some of this s
860.
by mrkurt 7y ago
That would work great. Even something like FaunaDB would work really well.
861.
by mrkurt 7y ago
That's actually the plan! The Ruby Graphql gem lets you resolve queries directly so the app can have an "edge" or "core" mode and make the right call (edge over HTTP, core direct to DB).
862.
by mrkurt 7y ago
These are great. IoT seems like a thing you could do but that's a really specific use case I hadn't even considered. The second example would be interesting to try. There's no real limit on VM lifetime or outbound connectio
863.
by mrkurt 7y ago
The containers on fly.io can go up to 8CPUs. So you can do a lot of computation for images, video etc. They can also accept any kind of TCP traffic (and we're trialing UDP), so lots of interesting network services. This is especially i
864.
by mrkurt 7y ago
Good call. We have a neat visualization we want to build to show what's happening, but we don't need to wait on that to make it more obvious.
865.
by mrkurt 7y ago
This is actually one of my favorite things about Hacker News. I end up redoing landing pages each time we get comments like this.
866.
by mrkurt 7y ago
Game servers are a great example. What's interesting is how many different kinds of apps need game-server like infrastructure: https://www.figma.com/blog/how-figmas-multiplayer-technology... We've been talkin
867.
by mrkurt 7y ago
That's exactly what happens when you disable Docker. We default to local builds because it's a little more secure. It's usually faster (our remote builder isn't great at caching layers yet). It would make sense to be abl
868.
by mrkurt 7y ago
The "simplest" gains come from adding an in memory cache, we include Redis for this and some apps work really well just leaving the DB where it is, caching aggressively, and running close to users: https://fly.io/d
869.
by mrkurt 7y ago
We actually have servers in North America, Europe (including Frankfurt), and Asia Pacific. The complete list is here: https://fly.io/docs/regions/#discovering-your-applications-r... Building an anycast network is
870.
by mrkurt 7y ago
Yes! It's really common to use airport codes for regions. There are some really strange ones, too. Someday I want to have a datacenter in San Carlos so we can have a SQL region.

More ›