Kong Kong Tutorials, 8 Nov 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Application Observability With Kuma Service Mesh

Description

The more services you have running across different clouds and Kubernetes clusters, the harder it is to ensure that you have a central place to collect service mesh observability metrics. That’s one of the reasons we created Kuma, an open-source control plane for service mesh. This tutorial will show you how to set up and leverage the Traffic Metrics and Traffic Trace policies that Kuma provides out of the box. Try Kuma: https://kuma.io/

▬▬▬▬▬▬ TIMECODES ▬▬▬▬▬▬

0:00 - Intro
0:36 - Kumactl Overview
1:10 - Kuma Mesh Dashboard in @Grafana
1:48 - Traces and Logs
2:37 - Kuma's Out-of-the-box Observability Dashboards
3:35 - Fault Injection Policy Scenario
5:04 - Using Jaeger Dashboards for Analysis
6:41 - Conclusion

▬▬▬▬▬▬ LINKS ▬▬▬▬▬▬
❏ Read the Automate Service Mesh Observability with Kuma Blog Tutorial: https://konghq.com/blog/service-mesh-observability
❏ Watch more recordings from Kong Summit here: https://konghq.com/kong-summit/2021-videos

#ServiceMesh #KumaMesh #Grafana #Jaeger #KongSummit

A

The more services you have running across different clouds and kubernetes clusters,.

A

The harder it is to ensure that you have a central place to collect service mesh.

A

Observability data., that's one of the reasons we created the kuma,, an open source control, plane.

A

For service mesh. in this demo,, you will learn how to set up and leverage traffic metrics,.

A

Traffic trace, and log policies that kuma provides out of the box. kuma control command line, utility,.

A

Allows you to install different stacks that can support and collect different metric information,.

A

Tracing information, and logging information. so,, if you do kuma control, install logging,.

A

It will install loki.. If you do kuma control install tracing,. It will install jaeger,.

A

And so far and so on. for the sake of this demo,, we already pre-populated our service mesh,.

A

With our service data., you can learn how we can install and enable all these things in our.

A

Other demos that available in our book. so,, let's take a look on probably one of the most important.

A

Dashboard here, called kuma mesh. kuma mesh has a service map,, a component that gives you overview.

A

Of your service mesh., you can see different services that are connected to each other.

A

Each individual node of this graph will represent a different service.. We also can learn from this.

A

Graph is, rate of requests, how many quests the service is issuing. and also,. We can learn about.

A

Ratio of errors. traffic, information, error, codes, and the latency numbers also represented here on.

A

This dashboard. so,, you can have a full overview of what is going on here. so,. Let's explore some.

A

Of the information that available., we will start with some traces, and, let's analyze, some traces.

A

That happening from the front-end application., as you can see, from the cascade of the calls,.

A

That our client issues to different services or front-end. calls, recommendation, recommendation.

A

Calls, advertisement, and so far and so on. also,. We can see logs from the different services. so,.

A

Let's pick up a logs from front-end application., one of the things that we can see here in the.

A

Logs,, so we can learn from particular container that the sidecar container that includes some of.

A

The . we can get actual information about traces, and we can drill down in the jaeger panel, and see.

A

What is going on, and how the services are interacting., so each individual,.

A

Our request will include some of the trace id information.. Kuma comes with a number of.

A

Useful dashboards out of the box. so,, the first one, is the kuma-cp that shows the status,.

A

That of the control plane, and multiple system administrators., we can use this to understand.

A

What is going on, if the system is healthy,, how the communication's happening between different.

A

Nodes of service mesh and so far and so on. next one is the kuma dataplane dashboard. in the.

A

Kuma dataplane dashboard,. We can learn how many dataplane sidecar services are running.. What's our.

A

Stats there, and so far so on. we can learn this per particular service.. Next dashboard is kuma.

A

Service, that allows you to see different traffic information that happens on particular services.

A

We can see for particular mesh,. We see this advertisement service, see some latency numbers,.

A

We see some ingoing, outgoing traffic. also,, there's a useful dashboard that allows you to.

A

See traffic between particular service., so in this case, we established front-end and advertisement.

A

Service, and see how the traffic flows through between those. so next,, let's use this dashboard.

A

To investigate some of the problems that might happen in our system. for this,, we will create.

A

A fault injection policy that will specify certain interesting parameters in our system., specifically.

A

Here, we will introduce some of the delays between front-end service and advertisement service,.

A

And every 10% of request will have one second delay. so,. We will apply this, and let's take a.

A

Look inside our kuma mesh dashboard, how this will affect overall application, performance.

A

As you can see, this, from the latency panel,, we start seeing that the latency, it starts piling up.

A

Between our front-end and advertisement, service. next,, let's introduce another type of.

A

Fault injection that will introduce some of the errors while calling card service,.

A

Will be calling checkout service. so, every 10% of requests will end up with 404 error.

A

Let's apply this policy, and see how it will change overall picture of our application. now,.

A

We can see some of the nodes start accumulating some of the errors, and our front-end service.

A

Have some errors? Because there is communication between different services start getting some.

A

Errors. so,: let's take a look where you're using service-to-service communication, to see what is.

A

The traffic between our card service and checkout service. we can see on the status code, graph,.

A

How many status codes we have with 404, and how many of those are 200. next,? Let's do something.

A

Interesting., we will remove all this fault injections that we have here, and we'll introduced.

A

One,, but we're not gonna reveal this what it is. so in this case,, we will show you how you can use.

A

These dashboards for your root cause analysis. so,. What do we see here on the service map,.

A

Is that we start getting some errors in the recommendation and payment service.

A

So, we will see that payment service is starting getting a lot of errors. so,.

A

Let's investigate this by looking into traces. we're going to explore panel,.

A

Switch to jaeger, and let's find these calls from the front-end to whatever service, where the status.

A

Code is 500., we'll use this last five minutes, and let's analyze the first trace that we have here.

A

Immediately, from the jaeger panel,, we can see that the front-end called, card service,.

A

And card service, called payment, service, and payment service called, recommendation service.. We see that the problem might have inside this recommendation. Services,.

A

And we know the exact code that recommendation service throws in. so,, let's put additional filter.

A

For recommendation service, and see what else we can find here. inside this panel,, we can see.

A

Recommendation service have some of the errors, and let's see, communication between front-end and.

A

Our recommendation service. so,. Clearly we see the number of errors here is starting piling up. and.

A

That's how we can use this panel to investigate and find the problems in your system.. Don't forget.

A

To check other demos in our virtual demo, booth panel,, where you can learn more things about.

A

Cloud-Native connectivity, service, mesh, and kong. and with this, and as always, have a nice day.