Home
Contribute Contact Us Browse all meetings
Home Contribute Contact Us Browse all meetings
Kong / Kong Builders Livestream Recordings / 8 Jun 2022
Kong / Kong Builders Livestream Recordings / 8 Jun 2022
Previous Meeting Next Meeting
⏯
Add meeting Rate page Subscribe
youtube image
►
From YouTube: Kong Builders - June 8 - Kong plugins you don't use but probably should

Description

Kong Builders is the livestream series that takes our developer-focused toolsets and puts them on display in the best venue possible – building applications and connecting workloads.

This week, Viktor Gamov will review some of the Kong plugins you may not be using but probably should, including #Kubernetes #Ingress controller and plugins on #k8s.

#KongBuilders #developer #livestream #livecoding #KIC

A

So.

A

So.

A

All right, uh hello and welcome to this episode of cone builders welcome to uh kong, and uh previously I was saying, welcome to call youtube channel, but I know that right now the stream is actually going to many places so welcome wherever you are youtube, linkedin twitch um and we are streaming from everywhere, write down in the comments. Where are we coming from like geographically, and uh I would know that uh what kind of platform you coming from um my name is victor gamma.

A

I work as a developer to get here with guang, and hopefully you will spend next hour with me learning something new about conch or maybe not learning something new about it or maybe I will learn something from you all right. So, let's, let's get this, um let this thing going and let's see if you will be able to see me and the and clearly hear me, I see uh marcelo from linkedin coming to us. um Welcome welcome great to uh great to have you here.

A

um Let me know how you, how is everything I'm monitoring on my site and see uh the uh looks like a stream is going strong um and uh hopefully you also enjoying this- um the uh the home builders? Okay, so cone builders, uh what is gone? Builders, the first time here, uh if your first time here for coin builders, uh this is the show that we're trying to do weekly bi-weekly once in a while, but like it's always on wednesdays uh and and wednesday we're talking about all things kong.

A

uh This is unscripted unprepared, live stream where uh we're trying to build things because of co builders, but sometimes we break things.

A

um Usually I break things and I fix the things after after the fact, which is ridiculous, but hey that's the fact so uh clunk builders, you can always uh subscribe to two or a few fresh episodes if you're not interested in the stream by the way on the colin youtube channel, you can find the demos that work all the time and you don't need to sit in here and listen to me for one hour.

A

Okay, so um the I see that uh marcelo comes from brazil welcome and I see rahul comes from from india. Also welcome. It looks like um now. The time frame that we picked up for this show is actually very friendly for many time zones, so people can join uh all over the world. So it's great um it's great to have you around here so now, uh before we jump into today's subject, so I will show you where you can find all the episodes.

A

um There is a uh calling youtube channel, like all uh things that I'm doing are linked there majority of the things. I also sometimes uh write blog posts, but we'll talk about this um and goes into here, and you can find this. uh The con cone congress controller. No, no coin. Builders live stream, so you can find all the episodes all the previous episodes.

A

um I highly recommend to check out uh the previous episode where my colleague cat actually broke down how you can deploy kuma, which is, since you have sandbox project, that we at contributing actively uh how you can run this in ecs. I think it's it's a pretty pretty exciting topic. If you in this, if you in this type of jazz, I highly recommend to check this out from the last time and uh also like hit that like button and uh encourage cat to do more content around service mesh okay.

A

So uh where are we here?

A

Okay, uh thanks for uh complimenting uh my uh t-shirt, it's uh actually it's exclusive quon, gingras controller t-shirt. um So we actually doing the um kind of ruffle. So if you um sign up for another show that we're doing it's called comcast um and uh you will be able to uh to win uh to uh will you will be able to win um the short.

A

So if you will subscribe for this, uh this email- oh I'm, not showing this okay, so we have another show that uh that we're doing it's called comcast and we actually can do a raffle. um So you sign up there and uh usually we send uh winners, some of the swag so holding this cool swag with the uh quan king dress, controller t-shirt, for example, for example, but also um if you are in the kubernetes type of jazz.

A

I highly recommend to listen our previous episode that I recorded with jessica dean from microsoft, and we talked about bridge to kubernetes, which is a pretty cool tool that allows you to debug your kubernetes application in real time. um So that's um that pretty much. Oh! No, the one last thing uh actually two things that I just uh I just realized that I didn't mention so first thing is that um calling summit con summit is upon us, so you can register for con summit.

A

You can get the early bird tickets, uh so let me remove this one from the stream yeah. um You can uh register for early bird um tickets, get some some discounts, so kong summit is going to be a pretty cool event where we will talk about all things: technology around the corn, but not only conch, but we will have a speakers around the industry. uh Talking about uh the cloud connectivity apis, um some some cool use cases how they use quant or somehow can use cloud native technologies, and things like that.

A

So I highly recommend you to join us in september in san francisco. It's going to be a very, very cool event and yeah, so the prices right now uh sales and in 10 days, if you it's a two-day conference, highly recommend to uh to join us, but there's a there's, a hack, some people asking victor how I can go to conference and not pay for for the conference. So that's what I personally use. uh That's my uh that's my personal you! Why go to conference?

A

If you can speak at the conference, so- um and you can go here- find this page about consignment call for papers and you can submit your talk about your api and innovations and all things uh cloud connectivity. uh If you will be chosen to present you don't need to pay with conference. I think it's a uh it's a huge uh huge bonus. It's a huge win, all right, so that's the um that's uh all announcements uh that I wanted to.

A

um I wanted to show right now but like if you have some announcements, uh send them me to uh dm uh in in you know: twitter, uh you can you can find my twitter here, it's always in in my um a lower third that is down below um you can send me dm on twitter and I can read your advertisement for this live stream. We have a coverage across multiple platforms but anyway, so, let's, uh um let's enough with all this jazz.

A

um So today I wanted to talk a little bit. You know I would just try to experiment with the format, and I tried to bring some of the things that um I found interesting. This is how this show is supposed to be doing like I found something interesting, so you will find this something interesting, so um I was doing a presentation in um in um let me switch to uh to my channel.

A

So I did this presentation on uh in dc copenhagen uh last week uh was cool conference in uh it was cool to visit some some northern countries, and I was talking about the all things that available to you as a developer when you're running the con as a ingress and what kind of things uh you can do this, but because of the format of the conference, I couldn't possibly put everything I just like scratched the surface, but explaining what's ingress you can you can check the slides and the video will be available also very soon, and I talk about ingress.

A

I talk about http, gps and after that, how we can use um or deploy services that not uh maybe um available as a http or l7 like jrpc or gcp or udp. So I showed all the things the how you can do this with phone.

A

The regular uh listeners and uh regulars in this channel also know that I did some of the episodes of kong builders uh on this subject. We talked about tcp ingress gdp, ingress all this all these cool things, but I was playing around with some of the things and I um I found that there's some plugins that I never use, because, usually you know if I want to do kind of like a one-on-one talk and talk about the ingress controller.

A

I have just like a small small set of plugins, and maybe many of you already have seen those multiple times rate limiting some of the key off type of plugins. If you haven't seen this again, there is a video on the youtube channel. You can watch this.

A

um I explain how those two plugins you know work together, but I also found the plugins that I never kind of used, but they kind of cool or or useful at least so, um and maybe today would be one of the episodes where I will start um talking about those plugins start talking about something like gems or how my friend and colleague picked up this morning um kind of.

A

um Could you pick so he shared some of the um past when he worked at the blockbuster, which is kind of like a dvd and vhs renting company, and there was a corner like when the staff staff pick uh can pick up some of the interesting things some interesting movies so consider this is the peak of my uh plugins. If you like, this format, uh write down in the comments, I will continue to do this, but so and we will jump in as always before we start.

A

Let's see.

A

There's nothing here installed, so we're gonna do.

A

We just go ahead, install clonk in uh namespace kong. We're not gonna, do any particular version. So that's why we can install like fresh as always yolo um and uh something might not work. Something will break definitely will break, but I will see so uh once we'll install. It depends on the uh depends on the speed of my internet today. uh We're going into the space quan uh conc is starting up we're using the latest and greatest bits and we'll see if we can get the ip address that we usually have.

A

So in this case I don't need to.

A

Can I do this? Can I do this like this? Oh okay and let's see what's my proxy. No, it's a different one. So for simplicity, give me give me a quick second. I will try to register this external ip address for for my application. So in this case we don't need to mess around with. um We don't need to. uh We don't need to mess around with ip addresses, and things like that.

A

Sometimes when I, when I do this kind of demos, google can assign the same ip address for my load balancer. But this time I was out of luck and my ip address that I already have um just just disappeared. Okay, so I will quickly update my dns um and we will see how fast it does. Oh you see. So dns works great. It's it's always great to have a good provider. um If I do proxy me, okay, so we're hitting this latest and greatest we're getting this response, which is cool.

A

That's what we usually like to have. um The first thing: that's um uh first plugin that I want to try to run. This will require a little bit more setup and hopefully you will enjoy seeing me set up this one.

A

So the for this plugins, so if you interested in plugins first thing you need to go is to plugin hub docs.conc, dot hq slash hub. This is where you will go and learn about the plugins.

A

In this I will for today I will try to focus on some of the you know: free plugins, plugins, that's only available for uh for open source column and and so far so on. um Maybe in future we will look on some of the enterprise, great plugins.

A

uh That would be also cool because those are having some of the some cool capabilities um and, let's see I wanted to start with something simple, a pro prometheus, so I want to have this uh uh start collecting some metrics from uh from my system. So that's why um I will uh use this uh this parameters uh plugin. So as always, you can uh you can uh you can go and find this in our in our plugin hub and uh we're going to be installing this in uh in in in kubernetes.

A

So one of the things that I don't know, if you have how many, how often you spend time in calling documentation- but we just recently did very small but very nice uh introduction, so all configurations now not scattered around this document, but like uh very nicely and compacted uh organized. So, for example, I want to enable this plugin globally, and I want to enable this in kubernetes. How cool is that? Just you know, just click around drill down. I think it's cool and we're going to be using this this this configuration.

A

So I do have this things uh prometheus plugin thing already available in my editor. So, um but for this I need to have some of the configurations that are available in in quang itself. Specifically, we need to enable metrics and we need to enable this. um The portal notation so prometheus will be able to scrap it. um Those values I don't need for now, so I will just uh I will just uh where's the I will edit those and enable enable those annotations.

A

So in this case, let me move it here.

A

Yeah, so I move this into my uh into my directory so um and when I just do helm upgrade uh with values. Where do you see ammo.

A

Again, enabling those annotations um allow to this plugin to scrap this or not the plugin plugin will be um enabling the traffic that will be hitting your ingress and the pass this information down to prometeus, but we don't have prometheus and we don't have a tool that allows us to visualize it.

A

So I was thinking about this and you can actually find the full-blown the blog explaining how this work, if you go to a application, performance, monitoring and there's a full blog, explains how you can set up prometheus and grafana for for the application, um this blog uses, uh kind or kind kind, um again write down in the comments, if you, um if you know how to pronounce this correctly, um the um we're not going to use it, I'm you know what I'm a fan of the uh running down the real hardware and like a burning money um on the on the real, like regis cluster.

A

So we're we're not going to do this locally. um A couple things that we need to have. I already uh have those um help charts for prometheus and grafana. Already in my uh in my uh so I'll. Just do helm repo upgrade update.

A

So, as you can see, grafana is here: prometheus community is here as well, so that's we already skip so in order to install, we need to enable some of the values for prometheus, so we should say for 10 seconds uh it's how often these metrics will be collected, so it should be uh more than enough. So we're gonna do this. uh We can install this one.

A

um This blog post uh uses particular version. Like I said in con builders, we we don't use, we don't use the uh the the um we're not we're not using like a safety belt, so we just use like whatever version available and after that yolo, so uh we're gonna be installing prometheus uh using a community chart uh in the namespace monitoring will create this namespace and we're going to be using prometheus yaml as our values.

A

So we're going to go here and go into that's a key9s, my favorite tool um to show you kubernetes jazz in in the comment line, or rather it's not common line, it's kind of like a text based interface um and we're starting bringing all these things here. Okay, so um we're not going to be accessing this directly from from our application.

A

We're going to be using grafana to show all these metrics now, and the next thing is that we need to install grafana and uh in this blog post uh also, I use the values that we need to install specifically. This is gonna, be our primitive server. uh Let's see if it's a mum.

A

With the these, primitive server should be available within uh within our kubernetes cluster, uh by a particular dns name, yeah, so primitive server. So it's it is available from this service name. However, since we're going to be running graffana in the same namespace, we don't need to specify the namespace.

A

We don't need specified service cluster local, so it will be available based on this um service name, um if you're running this in the different name, spaces uh make sure that you're putting the full full qualified path, all right um so uh and the rest of the stuff. Oh okay. So that's um that's actually neat! So if you want to use something from, um if you want to use grafana dashboard quang actually has pre-built a grafana dashboard that you can use in your application.

A

So that's why, uh when we're setting things up, we specify okay, we want to have some default dashboard and the graph while we install it, will fetch it in installed in your installation. So um with this, let's go ahead and uh install this one.

A

Burning burden again, yolo we're not using any versions. Let's see how it goes. Let's see how far we can go with this while it's installed. Let me check. um Let me check any comments that we have here so far, so thank you. Taryn for the posting, some of the some of the links uh tearing is helping uh with uh running the show uh on the on the background always uh appreciated, and uh it's always uh great uh great, it's great to have um the the wing.

A

Lady, like you in my um uh conveyor, show okay, so we do have a few more people. uh Artio welcome uh it's great to have you here. I hope you will stick around until the end. um Okay, so rahul is asking. Can we cover some aspects of kuma as well? um Not, today, though, you go and subscribe to, where is it on our youtube channel and you can find all things around the?

A

um um Let me see: okay, that's wrong button yeah, so you can go and subscribe to our youtube channel and you can find uh some of the videos around kuma, but also um write down uh in the comments. If you want to learn more about kuma and what exactly you want to learn because there's tons of things we can talk about this. For example, in a few weeks I will be talking um in in java community uh about uh servicemen for java developers, or I will be talking about zero trust security with service mesh.

A

So all these things, um if there would be recording, I will post those, uh but you can always ask me- and I can do a special episode on this or cat now, our my in-house, my colleague and my in-house expert in the service measure.

A

Okie dokie, okay,.

A

Okay, so while we were talking while I was talking- and you were watching um they- um I don't see, give me a second um anyway yeah, so um grafana is up and running and uh to get a password for grafana uh we're using this metric command. So what it does um it'll get the secret from grafana and getting admin password and decode this on base64, this very safe password, and I think this password is actually generated like all all the time when we install it so we're going into grafana.

A

We want to do port forwarding port 3000 um for those of you who uh marvel fans, uh love you 3000 uh write down in the comments. If you get the reference.

A

Now and we're going into uh localhost.

A

And we're going into localhost.

A

It will ask me admin password and password. We already have it here.

A

We're logged in and uh yeah dashboard is up and running, but there's no data, there's no data uh because first we don't have anything. We don't have any any service and we don't have a plugin enabled so um the thing that's we do have uh now uh prometheus and grafana. So we're going to go ahead and enable prometheus plugin. So I will just go ahead: um it uses uh the custom resource definition, customer source definition and uh the the clone congress will be using this guy, so going in just click apply.

A

uh Essentially what does apply does is just a nifty feature of of jetbrains id that allows me to do. Koop control apply minus f with this file, so cluster plugin uh created. So let's take a look. What it's look like in our.

A

So configuration is updated, blah blah blah. So that's the output from um from ingress controller, and we see something with uh plugin. Let's see if we have a a plugin is available, get plugin schema. Okay, prometheus is up and running. So, let's see, um let's see if, as always, we um um the way how we can generate some traffic, how we can generate some traffic, um I'm using uh many of you know this. I'm gonna be using insomnia um and a new http request.

A

It's a base url and my base url for this environment. I guess, uh would be cornproxy.me yeah. So if I will send no route, that's totally expected. However, we want to see if the if the date they start actually going in uh so last five minutes. We will refresh um refresh some data and uh looks like um uh we start getting some data from uh from from law worker, but there's nothing here. So there's no!

A

um There's no yeah, there's a few connections here and there, but there's no service because we didn't deploy anything- and um many of you know that I am um uh have my favorite application. The code service, that junior is code source about chuck norris. So, that's why we're gonna gonna go here and we will go do chuck.

A

So we have a deployment which is go ahead and apply this deployment. We create service.

A

Apply this one- and I don't have this plugins installed, so I don't need to enable those and uh chuck is going to proxy with me. Okay,.

A

Now so I should be able to get back to my uh chuck norris. Let's see if it's a it's working, yeah, it's working, so ingress was successfully created, we're getting the response uh from uh from correct from correct from correct port, which is always nice now. So uh let's generate some traffic, and in this case I will say a repeat on interval, 0 5 right and we should get. We should start getting some of the.

A

So when I refresh here, I should be able to see a new service will appear. Let me reload this one, um my my new service. Now um I start seeing the data from my service and since I'm running this into say, I will show you deployment.

A

Replicas 2, so I'm running this server and the the qual is a lot bouncing a lot balancing this. Let me stop this for a second, so right now um how I can know if it's actually does whatever I said it does.

A

I have a two instances of my service and calling does round robin between those two services, and if we go here in my namespace with my defaults, two instances of my code service see it's something different. It's nothing to do with this guy um check quote service. Has uh this two things at the end 6g m rx in hdcgm, so if I will go and click insomnia, I will getting the response either from this pod or from this spot.

A

So that's why we are seeing here um in this two services in my dashboard all right, so um I will be able to see your request rates. I will continue to.

A

Repeat, on interval just do every five seconds.

A

um So far, so good uh works great in the check this out, I have a ascii art of my con inside my console anyway. um Okay, so the first plugin. That's that's what we broke down right now, so we took the uh enable prometheus and start collecting metrics out of the ingress, so um it's very easy to do with uh with quant and start getting some of the relevant data for your application. Now. So next thing that usually um uh let me get back to this one.

A

uh Usually people um like to show the rate limiting plugin, because it allows to kind of like a push back uh in more or less um kind of like intrusive way, um you're saying. Okay, so don't don't touch me kind of like stop back off, um but also there's another plugin that allows you to do this configuration, but in um more kind of like a more you you um if you feel that someone is hammering your your upstream service, what about you give them cash response? So, let's, let's do? Let's do that?

A

So if I will go and find something cash.

A

There's a bunch of plugins but, like I said we interested in free plugins today, so we're going to be using the proxy cache, so proxy cache oh switch back to uh to a screen.

A

The proxy cache is the plugin that we're going to be using today, so we're going to enable this globally in kubernetes. That's going to be our that's going to be our configuration today here now.

A

Now now file is a proxy.

A

So, let's, uh let's take a look. What do we have here.

A

So a return response code for hand for get in hand because you probably don't want to cache post request content type application, json cache ttl. Let's take a look: what's the ttl in this case means and in seconds 300 seconds. Let's do this something like uh 60 seconds, so our cash will be updating here, strategy um strategy here.

A

So only accepted value is memory. Okay, um hopefully, let's see if we can do there is the advanced caching plugin that has global kubernetes and it has strategy.

A

I think it will be using.

A

Strategy memory um redis, so it has a red as a as a configuration here, um one of the blog posts that we posted recently from uh from one of my uh from the our clone partners and from my uh the past companies I used to work in the in this company uh called hazelcast uh caselkas just released a plugin for for uh for calling that you know. Instead of redis, you can use, cancel cast for storing rate limiting stuff, so that would be cool to see also um caching functionality for hazelcast.

A

So um if you are um into advanced rate limiting, actually you know what um I can break this down. Maybe one of the few episodes. If this uh this calling builders episode will get, I don't know like 150 likes. Let's do this and I will break down some of the advanced features.

A

um Okay, so we're going back to uh to my conversation so for this um we'll call it global proxy cache uh we're gonna be using for ingress for this guy, and uh we will enable this also. We can enable this on ingress. I believe so, let's do this one. We don't want to do this global. Let's do that global, because.

A

Whatever, like, we can do this next time, we're already 30 minutes in all right, so object method, labels, type of string, so we're looking into something typo. um Can it unmarshal ball to go to struct labels yeah, of course we're gonna? Do this.

A

Okay, now we just do apply, and here we go now we should start getting interesting responses here.

A

uh Let's see, stop response and cash status hit and immediately, as you can see, application was able to get upstream. Latency went to zero and because the proxy will be responsible from here, let's see where is it the response? Where is the where's? My check noise.

A

Preview of response you see- and I think for the next um repeat- on interval: let's do every five seconds or two seconds. I don't know why yeah um for the next uh for the next minute, we should be able to get um a same response from from application, and since we have a matrix already um somewhere here, so let's refresh it um we're writing this for the at least 15 minutes now so total request per second and uh requests per um per one half request interesting. Why? It's not why it's a half of request?

A

um So essentially what happens here? So we not hitting this directly ah two requests per second, meaning that every half a second we were submitting this request, and now it just went down because we just need to hit this one once in a minute, um so our um caching plugin works. So here we go the the caching plugin available uh for free. You can just enable this um and you will be able to see how you can.

A

Some of the responses doesn't need to be a real time. You know the the this uh this particular um use case where I'm posting the quotes from chuck norris. Those requests doesn't need to be. You know real time so responsive um another. That's that's! A proxy cache go check. This out, um go check this out proxy cache and uh we're gonna to look to um in another um interesting another interesting point.

A

So let me quickly check if you have any uh comments here, I'm wondering um people, let's, let's make some noise, how many people are watching us everywhere, I'm just curious. uh How many people are there? Still there write down in the comments and hey I'm still here, I'm enjoying this? That's uh that's super cool show. I am happy af. Maybe you even have some questions that I can use as a breeder.

A

And um let me see.

A

I mean I don't I don't see how I can tell how many people are watching.

A

Oh, I know I know I know I know, but I'm not gonna do this right now um or can I view analytics yeah, so uh we're using this uh tool called restream and we see um lots of statistics and some of those statistics are in the real time. So that's that's pretty cool, so I know that few people are watching on youtube and people linkedin just dropped to watch. Probably they got back to work because you know you can use linkedin for work all right.

A

um Let's get back to some of the plugins um plugin that you can use to kind of like push back on. Certain workloads is.

A

Sometimes you can have uh bots like uh this is computer computer software that will be going around internet and trying to hammer your system or scrap your system uh for some useful data, and you don't want to do this, and this is where the plugin uh can come: handy, yeah, super cool and a small plugin that allows you to put extra extra spice on your uh on your level of protection of your apis, so what it does so um it actually analyzes um um we can, uh since, since it's open source and the source is available over this plugin, so you can always go and check to see what this uh plug-in does.

A

So, essentially, this bulk detection plug-in will analyze it's a bunch of different um regex.

A

Regex regex that it will analyze and after that push is back. There's a oh google plus google, plus google, google plus um write down in the comments. If you still remember, if you're old enough to remember, google plus.

A

Google plus anyway yeah so um based on the user agent of your client, um you will be able to you- will be able to push back. So, let's see, um let's see how it goes. Okay, so, uh in order to enable this again we're going into our uh the portal enabled globally, we just enabled it globally. We enable this through kubernetes and just copy this. This parameter.

A

I have this already copied here and deny coin builders.

A

um Hopefully this will work so deny um if we look into the configuration. So it's a it's an array you can put uh uh coin builders uh chromecast, like your your favorite shows um you can say I don't know like what what what's the what's the facebook will already have uh instagram.

A

I don't know without not with no reason um and after that I will just go ahead and apply this now. So in order to test this, as we can see, we continue to hammering and the the cash status cash hit, 36 38 and after that, after 60 it will restart okay. So when I will do this- and I put some extra header so, for example, user agent and I said uh column cast.

A

And now all of a sudden are getting forbidden. um So with this um the the this is how you can immediately enable this one and it's it's still kong, but you know microsoft, internet explorer, crawler,.

A

I don't know, apparently it is a thing somewhere in the internet. So, let's see, um let's see.

A

Yeah, it was a some sort of some sort of crawler speaking about ip addresses. Okay, so that reminded me that you also can do things with ip address uh ip restriction.

A

uh We go and enable this globally enable this using kubernetes and we'll just do. um Let's copy to my id.

A

All this all this code, all these examples will be available in in in the demo scene, repository.

A

Plug-In yellow.

A

Again, I need to fix that. Okay, um how I can get my ip address, get my ip address.

A

Okay, so let's see it will give my mp address. Apparently this is my ip address, um we'll see.

A

Okay- let's, uh let's see if I will be able to restrict this access to this api using this ip address global ip restriction.

A

Enable, let's see if it if it works, okay, let's do this crawler. We don't need this um boom for beaten. Why is that? Does it show me why your ip address is not allowed? How cool is that um it's pretty awesome? um I would mess around right now with some of the different um I don't know like vpns, maybe and play around and see how this stuff working but hey.

A

That's uh that's how you can get this um and let's see if, um if our dashboard actually have something uh we're already doing this for 30 minutes, let's refresh it um our service, do we have a response bandwidth? What is the bandwidth? Let's see, what do we have here, ingress, aggress, so ingress? This is something coming in. Aggressors will be coming out. um It's a virtual machines, blah blah blah engine x. um That's the request response from um from nginx and all this stuff total requests per.

A

Second again, the majority of requests are are globally cached. So that's why we see this latency went down so um not to myself like um in order to insomnia, is cool and stuff like if I can do just um just do repeat on interval and just do like every half. Okay, all right! So we'll do this one!

A

My ip address is not allowed, so I should get a bunch of um I should get bunch of like. uh Can I get uh I'm interested in if I can get the error rates from the from the quant plug-in, so something that allows me to uh let's see prometheus in grafana, but I'm interested in, let's see uh just go to plugins3, prometheus and.

A

um What I wanted to say, I wanted to say a code like http code status, codes status, codes available as a metric, but I don't feel we have um let's go ahead and explore um so it says: what's the status called cone status code, so I'm switching back to uh to grafana and where's the grafana grafana grafana, and I will go ahead and explore whatever save dashboard. ah Okay, no, I just did some modification and wants me to save dashboard.

A

I don't wanna, no, this card go ahead, go away now, metrics browser. Let's see.

A

To do select metrics and to do we should have quang http status.

A

And that's the http statuses that we have what kind of statuses we have uh four or three, obviously because um forbidden and two hundred, so uh we start getting those uh let's see for the last 30 minutes and um yeah still still hammering still doing the things. So you see that before there was no like one so yeah those metrics are available, so you will be able to uh put this in your dashboard. Maybe we can play around the next time with this dashboard all right um interesting.

A

So the uh looks like a lot of people are watching on the linkedin. That's awesome. um Thank you for um watching us on linkedin, because my statistics didn't show uh many people watching which is uh which is uh uh which is super cool. uh Thank you for support all right, so um I think.

A

um Let me put this.

A

I will just do a quick, um quick wrap about the stuff what we were talking today. um Builders episode is zero: six zero, eight twenty 22., so we looked into plugins.

A

So we we talked about the first plugin was a prometheus.

A

And I showed you um how you can set up plus the metrics set up. I showed you how we can do this um go ahead and check out the prometheus plugin for from from prometheus hub. This is something that you will be.

A

That you will be able to find next thing is that we look at the caching.

A

Proxy.

A

And you saw a plug-in and you saw how we can use this proxy caching, the let's go back. Let's just do.

A

Proxy cache plugin um this. This is the one that um uh you will be able to use to cache stuff.

A

um I still cannot believe that we had google plus google plus anyway. So next thing is that we touch the on api bot detection, plugin bot, detection.

A

And the bot, detection and board detection plug-in.

A

This is some someone just messaged me: what does it mean? Connect uh connect compatible. um So this is a. This is a very good. uh This is very good question um so essentially connect. This is something that we're working on uh at quank. It's called um manage control plane for uh for your application um right now. The everything that you saw in uh in this particular presentations and the presentations that I do around quank and grass controller. We have this um conceptually conceptually.

A

We have this thing called um quank, ingress controller and in this particular case, when we're writing this inside kubernetes. The quan ingress controller um is a control plane for my uh data plane, which is quant proxy, so proxy, it's the something that will be serving this port. This is something that actually listens when I use you uh when I do this- and uh this is some some process that listens to this port, this is data plane.

A

That's responsible for for shipping, for for handling your traffic control plane is responsible for accepting configuration validating in the ship disconfiguration into my data plane. So in this example that I showed you for the last few episodes of chrome builders, we use calling congress controller again. A cool t-shirt is our control plane. Now you can also run quank as a data plane in your kubernetes, but use something. That's called connect as your control plane. The connect has like a lots of cool cool things.

A

There's nice ui all these like metrics things you can enable plugins disable plugins through ui. um I did a while back when I just joined the cool clonk. I did the the cone builders episode, how to deploy um um kong into kubernetes and use connect as a control plane. We are very soon. I am I'm super excited because I've seen the um internal demos and I worked with the new version of connect for quite some time and um I'm super excited. I just like every time.

A

Every time when the chrome builder is about to air- and I was just like going to my pm folks and saying hey- can I show this? Can I show this? It's super cool and it's not yet not yet wait wait a second!

A

We want to make this uh very nice and super polished, so uh we're gonna be talking about connect a little bit more but yeah, so some of the plugins, meaning that when they said um so, it's connect compatible, meaning that you will be able to enable this plug-in uh through um through connect control plane.

A

um That's the that's the that's the idea behind connect compatible thing, so so stay tuned subscribe for uh for a new episode of con builders- and I will be- um I will be posting those um and the last thing that I want to show you right now. So all these uh all my demos will be available here. There's a clone builders folder. That includes all the dates of previous demos, and you can find all the cool snippets there. So um before what what else? We did the bot detection, ip restriction, ip restriction.

A

Ip restriction, uh plugin and.

A

uh And I just do peer restriction, plugin copy this one. So that's my um that's my notes that that will be uh we'll be posting after the show. um Hopefully, you will be able to check this out um proxy yeah. So that's that's. Gonna, be like a set of plugins that we covered today like super super easy, but they will immediately improve your um immediately improve your um the workflow where's, this api restriction.

A

Can I can I not do this so, let's see delete this configuration yeah, delete this resource um and uh hopefully I'll start getting yeah, and you see this is the beauty of this uh thing. So my all the things with configuration that I'm messing around right now they don't touch uh data plane at all because uh control, plane, data plane uh connection are not standing in um in the in the path of your request. Your request always hitting kong and like we continue heating conc and immediately when we change the configuration continuous controller capture.

A

This turn it into something that uh quant will understand, meaning that uh quan can add me an api and after that, just submit this configuration. So I think this is a super cool and also I can do proxy cache. um Is there any like http response code for cached content, I'm interested.

A

There should be something like.

A

uh Okay, I can do let's see if I just do 204 um for this cached content. uh I'm wondering what this will say. So, there's still output here uh cancel request. um I need to stop and header and it's still 200, but why it's not.

A

Why not 204 response code configured, I'm wondering, let me see if I will be able to delete this and redeploy this, maybe something with the ttl. I think it's something with ttl.

A

Well, I mean like not everything should work as it's supposed to cache status bypass by press bypass um response code. Let's, let's see wait a second. This is interesting. It's a it's interesting uh proxy.

A

Proxy cache and let's see configuration uh default value so response code upstream response status, considered cachable, ah okay. So it's a response code from my application, so it's 200, okay, ah okay! So when I hit the uh can I do this response code for my uh for my plugin.

A

I don't know if it's a good practice to do this.

A

Yeah, so that's why it says bypass because I hit this. uh My upstream service, it returned 200 and we said, like 204, will be considered as cash response. So that's why um that's why? It is what it is. So another thing uh you learn: your uh now is cmis and now it's a hit for the first time it was a miss in the second time it was a hit uh because it was not available there.

A

So um the the most important lesson from today's episode is um read documentation, it's important, um so yeah all right, and uh I think I think we do. uh We do have the good stuff. For uh today, um you can always always should subscribe in our youtube channel, um write your comments and your requests for upcoming episodes, and with this my name is victor gamov and, as always, have a nice day.

A

You.