text
stringlengths
0
1.8k
[454.26 --> 460.50] And increasingly, some of our roadmap features take it to the next level with things like
[460.50 --> 467.68] EBPF-based service mesh, which I think service mesh is a really great example of something
[467.68 --> 473.46] where by running code in the kernel, we don't have to instrument each individual application.
[473.46 --> 478.54] And that's a big benefit, going to make things much simpler for people to deploy.
[479.52 --> 485.46] So does Cilium, I know that it exposes all these metrics and all this like visibility into
[485.46 --> 489.16] what is happening under the hood, especially from a networking perspective and from a
[489.16 --> 490.18] communication perspective.
[490.72 --> 494.98] But Cilium, what are the components in the Cilium product project?
[495.08 --> 498.38] I'm not sure how you want to call it, because obviously there's like the CNI and there's
[498.38 --> 499.32] other things.
[499.42 --> 501.46] What are the big components that make Cilium?
[501.46 --> 502.02] Yeah.
[502.24 --> 508.12] So when you run Cilium, you install a Cilium agent on every node.
[508.72 --> 513.10] And if all you want is networking capabilities, then that gets you going.
[513.40 --> 517.62] You probably want to start being able to see those network flows.
[517.90 --> 523.78] And to do that, you'd install a component called Hubble, which collects this network information
[523.78 --> 527.38] and the Kubernetes identities associated with it.
[527.38 --> 533.82] So if you look at Hubble flows, you can see traffic flowing between different Kubernetes
[533.82 --> 535.02] pods.
[535.32 --> 541.88] And then there's also a Hubble UI, which pulls that flow information, brings it into a much
[541.88 --> 543.60] more sort of human readable form.
[544.04 --> 549.04] So for example, showing you a service map and showing you how traffic is flowing between
[549.04 --> 552.04] these different Kubernetes services.
[552.04 --> 557.08] And perhaps where maybe there are issues, you can see the packets that are being dropped
[557.08 --> 558.48] within that UI.
[558.70 --> 562.00] So that's very useful in terms of debugging a network issue.
[562.40 --> 566.12] What about when it comes to alerting, monitoring, that side of things?
[566.24 --> 569.12] When there is a problem, you're being informed that, hey, there is a problem.
[569.40 --> 570.48] Is there such a component?
[570.70 --> 573.42] Or would you integrate Cilium with something else for that capability?
[573.72 --> 574.76] What does that story look like?
[574.76 --> 577.34] Yeah, you'd integrate that with something else.
[577.46 --> 582.88] I think a lot of people will push the flow data into some kind of sim, for example.
[583.40 --> 585.78] But I'm thinking about, for example, packet loss.
[586.12 --> 591.02] There's a lot of our congestion or lots of retries, whatever the case may be.
[591.10 --> 597.54] Is there a way to monitor or to consume the Cilium metrics, I'm assuming, and then have alerts?
[597.84 --> 603.68] So you can absolutely get the metrics into Prometheus or into showing Grafana.
[603.68 --> 606.94] Yeah, there's some beautiful screenshots.
[607.24 --> 611.10] And I think I can't quite remember where I saw them recently, but just this whole series
[611.10 --> 616.00] of amazing Grafana graphs that you can use to diagnose your network.
[616.20 --> 616.46] Okay.
[616.82 --> 620.14] I'm not sure whether you can tell by now that I'm really interested in trying Cilium out
[620.14 --> 621.80] for real in a production environment.
[621.92 --> 622.46] I really am.
[622.48 --> 624.28] And I'm trying to figure out what the components are.
[624.74 --> 627.88] So my next question would be, where would you recommend that I start?
[628.16 --> 629.36] Do I take the Helm chart?
[629.48 --> 630.48] Is there an operator?
[631.02 --> 632.46] What does the getting started look like?
[632.46 --> 635.02] So there's a few different options.
[635.42 --> 636.48] There is a Helm chart.
[636.80 --> 642.46] There's a command line tool, the Cilium CLI, which makes it as simple as installing the
[642.46 --> 646.90] CLI and then Cilium install and Cilium Hubble install.
[646.90 --> 647.68] I like that.
[648.40 --> 653.24] Does really make that getting started experience nice.
[653.74 --> 660.10] Also, if you want a helping hand, we're just about to start a series of weekly install
[660.10 --> 660.60] fests.
[660.60 --> 665.64] So the idea is to have a session with someone who's experienced in Cilium.
[665.86 --> 669.78] They're kind of guiding you through the process and it'll be interactive so that if people
[669.78 --> 673.06] have issues and questions, they can get help along the way.
[673.48 --> 674.72] So that's kicking off.
[674.80 --> 680.54] I think our first one is either this week or next week, but there's a new kind of feature
[680.54 --> 685.94] on the Cilium.io website to book your place on one of those install fests.
[686.30 --> 687.28] I love the sound of that.
[687.42 --> 687.58] Wow.
[687.62 --> 690.12] That's like the, I wasn't expecting for that answer, but that's amazing.
[690.22 --> 691.74] That's exactly what I'm looking for.
[691.82 --> 693.76] So thank you, Liz, for thinking ahead of time.
[694.68 --> 695.12] Thank you.
[695.18 --> 695.24] Sure.
[695.32 --> 696.20] This is perfect.
[696.38 --> 696.64] Okay.
[696.64 --> 698.14] I really love where this is going.
[698.64 --> 703.34] So I'm thinking of watching you code live, which is at the top of my list for this KubeCon.
[703.46 --> 707.56] It's like one of the must do for me at this KubeCon to watch you code live.
[707.88 --> 709.44] Can you tell us a little bit more about that?
[709.74 --> 710.94] Where the idea came from?
[711.24 --> 712.46] How do you intend to do that?
[712.58 --> 713.66] What are you intending to cover?
[713.98 --> 714.20] Yeah.
[714.28 --> 718.04] So I've done a few talks about eBPF programming.
[718.42 --> 722.60] There's lots of different frameworks and libraries that you can use and you can write your
[722.60 --> 727.34] user space code in different languages like Python and Go, Rust now as well.
[727.84 --> 731.54] My Rust isn't quite up to doing live coding in that myself.
[732.24 --> 733.88] What do you use for live coding?
[733.96 --> 740.78] I typically use either, Go is my kind of go-to language, but for ease of demonstrating a lot
[740.78 --> 746.78] of eBPF capabilities, I'll quite often use the BCC framework, which supports Python.
[747.20 --> 751.28] That's also very, I think, very easy to read in a live coding environment.
[751.28 --> 753.52] And occasionally I've done some that you see.
[754.18 --> 760.30] So because the kernel programs, the eBPF programs that you're actually running in the kernel
[760.30 --> 764.06] are typically written in C, can now be written in Rust.
[764.18 --> 765.78] So I'm going to have to up my Rust game.
[766.38 --> 773.12] But because the kernel part is often written in C, a lot of eBPF programmers are also comfortable
[773.12 --> 777.34] in that language and therefore writing the user space part in C as well.
[777.74 --> 777.96] Okay.
[777.96 --> 780.12] And what do you like to cover in those sessions?
[780.42 --> 783.78] Like, which is your, I don't know, step number one, step number two?
[783.84 --> 785.16] What do you tend to cover in those?
[785.44 --> 788.86] I haven't watched one, but again, top of my list, as I mentioned.
[789.02 --> 789.26] Yeah.