text
stringlengths
0
1.8k
[601.82 --> 603.52] And I'm wondering, does it still work?
[603.76 --> 603.94] Yep.
[604.16 --> 604.40] Yep.
[604.54 --> 605.42] So that still works.
[605.62 --> 606.36] We've maintained that.
[606.72 --> 611.62] That's not typically the production deployment, you know, because people are moving into repeatable
[611.62 --> 615.56] deployments and the Helm charts and code is config and so on.
[615.70 --> 616.32] Config is code.
[616.94 --> 618.18] But yes, that still works.
[618.66 --> 622.16] And I think that's still really important because a lot of people, believe it or not,
[622.28 --> 625.98] you know, Linkerd has been around for six years at this point or something.
[626.14 --> 628.62] You know, it was the first service mesh project ever.
[628.62 --> 634.48] But people are still coming into it fresh-faced, like never heard of a service mesh before.
[634.58 --> 635.64] I'm trying to understand this thing.
[635.72 --> 637.64] I've just learned Kubernetes, you know.
[637.68 --> 642.88] So there is a big audience to Linkerd every day where you're not ready to like helmet up.
[642.96 --> 645.32] Like you're just trying to play around with this thing and understand it.
[645.52 --> 646.50] So yeah, that still works.
[646.86 --> 649.20] How would you recommend someone that installs Linkerd in production?
[649.38 --> 654.10] So this is a very nice getting started, which I find very valuable, especially when I'm trying
[654.10 --> 657.22] things I love when tools are really easy to use.
[657.52 --> 661.50] And this is, in my perspective, one of the ways in which Linkerd is super easy to get
[661.50 --> 662.12] started with.
[662.32 --> 665.02] But how would you recommend that someone installs Linkerd in production?
[665.20 --> 670.74] Yeah, so what we've seen basically is people using Helm or Terraform or like tools that
[670.74 --> 673.90] allow you to do it in a programmatic and repeatable way.
[674.18 --> 676.22] And I think that's probably the best practice for production.
[676.70 --> 681.34] You want to be able to, especially if you're in the world of spinning out multiple clusters
[681.34 --> 686.76] or starting to treat your clusters as cattle and not as pets, you want those deploys to
[686.76 --> 687.46] be repeatable.
[687.68 --> 691.50] And you want to know exactly how things were set up when you come back to it three years
[691.50 --> 692.68] later, you know.
[692.72 --> 697.70] So you don't want it to be in someone's terminal window that they like, they closed their laptop
[697.70 --> 699.34] three years ago and then they left the company.
[699.50 --> 701.22] And now you're like, hmm, I wonder how this was involved.
[701.50 --> 702.66] So that's the best practice.
[702.92 --> 703.08] Okay.
[703.40 --> 708.20] One of the things which I've seen on there quite liked, especially when it comes to some projects
[708.20 --> 712.44] which can be a bit more involved to set up, is there's an operator which is just meant
[712.44 --> 717.48] to install things and then you apply a thing and the operator knows how to install itself.
[717.80 --> 723.10] Because then the thinking goes, the operator can also automate upgrades, which I think is
[723.10 --> 723.96] an interesting proposition.
[724.32 --> 724.38] Yeah.
[724.52 --> 729.04] So does Linkerd have something like that or is Linkerd thinking about something like that?
[729.12 --> 732.96] It's certainly something we've discussed in the past and I don't think there's a reason
[732.96 --> 734.08] why we wouldn't do it.
[734.08 --> 737.90] You know, the easing upgrades especially is something I'd love to do.
[738.20 --> 742.68] The upgrade to 2.11 is actually pretty easy, but going from 2.9 to 2.10 was painful.
[742.90 --> 745.26] Some of the configs changed and stuff like that.
[745.62 --> 749.46] I don't know that that would have been 100% automatable, but it would have been something
[749.46 --> 751.04] we could assist at least.
[751.36 --> 754.84] And there's other operations too, you know, that I think an operator would be helpful
[754.84 --> 755.06] with.
[755.36 --> 756.34] So yeah, we're open to it.
[756.80 --> 757.94] PR is welcome.
[758.36 --> 758.56] Nice.
[758.72 --> 759.30] Very smooth.
[759.48 --> 760.08] Very smooth.
[760.08 --> 760.40] Okay.
[760.88 --> 766.20] So the upgrade from 2.10 to 2.11, is it just apply the Helm upgrade?
[766.60 --> 767.38] Is that all it takes?
[767.60 --> 768.72] That really should be it.
[768.94 --> 769.76] We didn't change.
[769.82 --> 774.54] There's one or two breaking changes around the mechanics of some of the multi-cluster stuff.
[774.96 --> 778.42] But yeah, the majority of 2.11 is really additive.
[779.06 --> 784.06] And, you know, which again is a theme that we, you know, that we try and stick to with Linkerd.
[784.06 --> 789.34] So all of the policy stuff, you know, which was a new feature, that's all built on top
[789.34 --> 791.48] of all of the MTLS stuff, right?
[791.56 --> 795.66] And all that MTLS stuff is built on top of the Kubernetes primitives of service accounts
[795.66 --> 798.52] and mutating, you know, webhooks and whatever else.
[798.62 --> 799.88] It just kind of compounds.
[800.16 --> 804.16] And you get these very nice situations where, well, the moment you install Linkerd, I mean,
[804.20 --> 805.68] it's awesome that you can install it really quickly.
[805.80 --> 810.04] But what's even more awesome to me is that when you install it and you mesh your pods, you
[810.04 --> 813.24] actually have MTLS working out of the box there without doing any config.
[813.24 --> 817.38] Like, if you read that long, long MTLS guide that, you know, that you talked about, you
[817.38 --> 821.02] know, the vast majority of it's stuff, you know, like it's complicated stuff.
[821.10 --> 824.14] And then at the end, I'm like, but you don't have to do any of that because you can just
[824.14 --> 825.30] install Linkerd and it does all this stuff.
[825.60 --> 830.08] And that means that all the policy stuff can then be built on top of the identities that
[830.08 --> 831.04] MTLS provides.
[831.08 --> 835.56] So the cryptographically secured identities and, you know, it's all done in this zero trust
[835.56 --> 839.06] fashion where the enforcement point is at the pod granularity.
[839.14 --> 841.18] It's not at the firewall or the edge of the cluster.
[841.48 --> 843.10] So all this nice stuff happens.
[843.24 --> 843.48] Okay.
[843.84 --> 848.54] Do you have any dependency on something like cert manager or maybe a specific Kubernetes
[848.54 --> 849.08] version?
[849.48 --> 850.22] What does that look like?
[850.38 --> 855.30] So for Kubernetes versions, we basically try and, you know, support the most recent three
[855.30 --> 856.64] Kubernetes versions.
[857.10 --> 862.74] And, you know, often we'll have support for earlier ones, but it's not really the policy
[862.74 --> 864.34] is like, okay, most recent three.
[864.72 --> 869.10] Now, if you really have to, you know, do something with an older release, maybe we can make that
[869.10 --> 869.28] work.
[869.50 --> 874.32] In terms of dependencies on cert manager, there's not an explicit dependency, but one thing
[874.32 --> 880.56] you do have to figure out when you're running Linkerd is the certificate rotation, not of
[880.56 --> 884.10] the pods themselves, but of the cluster level issuer certificate.
[884.10 --> 889.08] We have some docs that have that automated with cert manager, or you can just remember to
[889.08 --> 889.36] do it.