Let’s Go Backstage: E2E IDP Tips & Tricks for Platform Engineers [Video]

Backstage is gaining wide adoption for platform engineering teams looking to build internal development platforms. After implementing Backstage, data has shown an improvement in 2X of code changes, and a decrease of 17% in cycle time making a huge impact on the business delivery pipeline for organizations. Backstage with its rich plugin ecosystem, makes it possible to get full troubleshooting and security coverage you need across your entire pipeline in a single dashboard and interface.

In this talk, we’ll take a look at why IDPs are gaining popularity, and Backstage has become the OSS tool of choice for building developer platforms. As well as demo how to get an inventory and scan your cluster, have a better understanding of your security posture, troubleshoot, and visualize the results in your customized Backstage dashboard through simple plugins. This talk will provide real code examples for how to DIY and build a full open source and fully secure IDP, with Day 2 operations baked in.

The following is an AI-Genrated transcript of the session:

Good morning! Today we’ll be talking about IDP tips and tricks for Platform Engineers. In this case, the P stands for portal and not platform because this is BackstageCon. I hope you’re all caffeinated even though after this opening that kind of revved everything up I’m pretty sure that we’re all awake.

We’ll be talking about the pains that IDPs solve, why use Backstage which is a question that I was talking about with somebody earlier on, and give you a few tips and tricks that we think will help you move along with Backstage. But let’s start with introductions my name is Oshrat Nir, and I am a Storyteller, a lifelong learner, a self-professed yogi who is more flexible in mind than body, and I’m also the developer advocate for ARMO and Kubescape. I am @oshratn on all the socials where I exist so feel free to catch up with me.

Hey everyone, I am Guy, and I’m a Solution Architect at Komodor, CNCF Ambassador, and I’m leading the Platformers Community. We are doing a lot of content around Backstage and I’m really excited to be here! For the last year, I’ve been working a lot on Backstage so I’m very excited…Okay let’s get started we all feel pain. Backstage solves some of that pain, what it does at kind of a very high business level that all of our bosses will like is that it increases software quality.

Well that helps us too because DevOps people don’t have to wake up at night and it speeds up deployment velocity which is again another great business KPI but if we were to think about our own personal pain; it reduces cognitive load and it improves our productivity which is a good thing and don’t just trust me – Expedia implemented Backstage and they were very very careful with their KPIs and this is the stuff they found after about a year of working with it, they were able to ship stuff successfully on day one they found that there was less context switching yay right nobody wants to do that and there’s increased reach of and usage of docks which is important because it promotes independence of developers and again it helps with the speed so why Backstage?

I can shut up and just show you the logos I mean everybody wants to be like one of these big companies they’re pretty successful we all know them but just in case it’s not about the love of the logos let’s start with it’s the fact that it’s open source right it’s open source it’s got a vibrant Community we all love open source that’s why we’re here apart from that it can be customized to your exact needs everybody in this room is a snowflake right you agree with me right and you can use themes you can use extensions and you can use plugins to make it yours the last thing is that it’s scalable and extendable so one of the tips that we’ll be talking about and returning to so I feel free to put it here is start small and scale up it will grow with you and you will scale the learning curve together with the portal lastly and I guess this is something that I found online when there was a question of why Backstage some people said that it’s really hard to implement and and it’s a big deal and why not build our own and you know not built here syndrome we’ve all suffered from it at one point the idea is this Backstage is a framework it’s a canvas you can make it your own implementing it initially the for the first time may be a little hard the thing is that with the new plugin system what happens is you do the heavy lifting at the very very beginning and then it starts to speed up and that is the important point and Guy will take it away from me now.

Thank you, thank you so plugins right everybody using plugins but now they’re becoming much much easier to work with so bad you mentioned and the previous called the new plugin system and and what’s the current state like everyone using today backstage knows that plugins require kind of heavy lifting you need to take the code you need to install the package sometimes the repo is not the repo of Backstage so you need to bring the code for some other repo and this is why they’re running the community repo for plugins and then you need to integrate that in most of the tutorials you need to add code to it and that’s kind of heavy lifting especially that we know that people that running Backstage most organizations are coming from non JavaScript background and and what’s really amazing about the new plugin system is that it allows you to really plug plugins you can take the image of Backstage you can add your own plugins to it without changing the base image you’re working on not the base image of the container but the base image is actually running and being deployed and and that’s amazing because now instead of creating a something that we thought to be extension it’s going to be a real extension extension point within the code and that’s going to change the approach with plugins and

And what we’re going to do today is that we’re going to show you tips of how you can use plugins and do a lot of things in platform engineering in order to elevate the new plugin system for that. So, one of the main things we want to talk about is continuous optimization, and we want to start with success. Like everyone using Backstage, we want Backstage to be successful. We want people to use it, we want our management to love the project and to invest more resources in it. And for that, we need usage because if no one is actually using the portal, we will not be able to actually show any success.

How do we do that? We provide some value in it. We hope that we are going to have usage, but the fact is that even if we invest a lot of time and resources to make it valuable, it doesn’t really mean people are going to use it. And that’s really interesting because one thing is actually missing in this diagram, and it’s the data. If we have enough data, we are able to actually understand how people are using the portal, and then we can actually prove our success.

Adding data to Backstage is easy. We are going to talk about three steps how we are going to build our analytics and data in Backstage, and how we are going to do that. So we’re going to start with the most straightforward way, and we are going to extend it step by step. The first thing is that we can integrate Analytics tools; it’s kind of baked into Backstage. You may need to install one of the packages, for example, the Google Analytics one, and pretty out of the box, you can get a Google Analytics dashboard of your usage of Backstage. And you don’t need to work a lot in order to get it instrumented, and that’s super valuable because we can get Google Analytics, they are perfect for optimizing websites and making sure you’re using them more and more and getting it into our deployment. So that’s the easiest way, it’s kind of pretty out of the box.

But we can extend it a little bit more. We got also the Insight plugin, and what the Insight plugin allows us to do is basically to bring usage insights into the Backstage portal. So let me show you a little bit how it looks. It’s a plugin made by Spotify, and it allows you to understand much easier how much people are actually using it. So, you know, like where the peak days, what happened during the weekends, and based on that, understand what is the usage pattern and how many users you have on a daily basis. So if you’re deploying a portal, you can go into someone and say, “Oh, we have 100 daily active users within our portal.” So becoming a product management metric and a successful metric that we can use.

And by the way, because it’s in the portal, and it’s not some in a hacky way in like go to that Google Analytics, we will have your credentials for that. They actually get it on their own with no RQ involved. And it’s not only us, you can also understand which plugins are the most useful ones, and maybe you launch a project to invest in the I know Sun Che plugin which we’re going to talk about later. People are not really using it, maybe we need to refine it, maybe we need to invest more time, and that allows us to actually take decisions to make it much better. Obviously, if you’re using the tech docks capabilities, adding docs into Backstage, if people are not really able to search them, they are kind of unuseful. So you’re able to understand how much search you have, and how much people are actually using that.

But it’s not the last step for analytics, right? Sometimes we do have events that actually triggered in the portal, but we want to add some attributes to it, and we want to create some correlations that are not intuitive, this is the most advanced level because then you need to go to the code and actually change the attributes of the events fired within the system. So we start from the basics, we are going to add Analytics tools, we are going to make sure that we have analytics somewhere, and get out of the B Dash Port. We can add the analytics plugin to make it much more advanced, and then we can also change if we need more data and then refine our data. And that’s how we can continuously optimize toward what’s going to be a successful portal and how we can prove our value using the portal.

And the next thing we are going to talk about is quality. The power that Backstage brings is that it allows us to set quality standards within the organization. I truly believe that quality is something that every engineer likes. We do like to provide and deliver high-quality products, but the problem is that it requires resources. In order to deliver that, we need a lot of resources and we need to invest time. But if we can bring quality into the organization and make it out of the box, we can actually leverage and elevate the standards and the quality standards within the organization, and that’s super powerful.

Now, we are going to do that by three simple things. In Backstage, we have the templates capability, which allows us to define templates and basically get them, and what’s interesting about them is that if you are able to define high-quality templates for the organization, people can create high-quality services out of the box. They can maybe get tests in them, maybe they can get ISO-based practices embedded into our templates, into our services, into our repos. If we can find an example developed by AWS, so they launched a plugin called OPA, and what they added to it is the ability to create and generate generative AI applications. Most of us are pretty early in the stage of creating high-quality AI applications, but how can we make them high-quality? Maybe we can take the AWS templates, we can use them, and then deliver our own high-quality AI application within the organization. So, anyone who wants to embed AI can use those templates and create high-quality standards applications. It can be for AI, it can be for microservices, it can be for a new thing developed wherever you like. You can use these templates and leverage high quality.

The next thing is scorecards. What we want to do is to reflect quality. We cannot control it all the time. We can create the templates, people use them, and then it’s kind of hands-off. Maybe removed, or maybe they changed the checks, but scorecards allow us to reflect and show users what quality their services are at. It can be in multiple options. For example, we can show them the test coverage of their services or microservices, we can show them the security, how secure their services are, also the incident, the reliability score. Everything about the scoring of the services will be reflected in them, and we can do that at multiple levels. We can do that at a team level, how good is my team? Obviously, no one would like to be on the lower score of the team scoring cards. We can also reflect that by services or applications and then understand where we are standing. That’s a good thing because people really looking for that. First of all, if you’re using your APIs, they want to know that you can provide high-quality APIs to be used. But on the other side, they do want to know that they can trust you, and this is the trust certificate for that.

And the last level of quality is Soundcheck. Soundcheck is a plugin that allows you to embed checks into your portal and make sure people understand what is tested and then get a certificate on that. So, this is a different plugin that can be used. And what we can see on the screen is, first of all, what the bronze level looks like and which tests are going to be tested. And what is useful about it is, first of all, we reflect the quality by score, and then we allow the user to understand what we have tested to define that score and what they need to do to make it much better.

Here, we can see, for example, what a bronze tier looks like and also how a silver tier looks like, and what they need to do, for example, they need tests to be passed on the CI, and from that, we can leverage, teach them, let them understand how they can get into high-quality standards of engineering. We see that use case many, many times, especially with SREs trying to enforce high-quality standards in the organization. We can see that from security teams trying to bring security into the microservices but they don’t know exactly how to do that, and that is a key capability.

So, let’s start from the first point. We want to integrate some scoring mechanisms. We can leverage scorecards or we can leverage Soundcheck. It allows us to go deeper into the check, how we can configure Soundcheck so it can be configured by code-based so we can define the check and make them run, or we can actually create them from the UI. It really depends. Obviously, we don’t want people to remove the quality checks of their own, and that can be a real problem, but it’s really easy to define, for example, a service without a README or without a catalog file, it’s going to be failed, and that’s really, really important.

And the third tip of the day is about creating a single pane of glass—a platform or a portal that people will be able to go into and actually see everything. This is really important because sometimes during the implementation, organizations tend to not actually complete this vision. They tend to overload the data into the portal, making it difficult for people to extract value from it. So, what we want is basically to add all the relevant data sources for the services into the portal and make sure that they get an overview based on all the data we can get. This is super valuable because we know that in every engineering organization, there are more than 10 or 15 systems that developers need to go through on a daily basis, and we want to integrate them into a single place so they can get an overview of all of them at once.

What is interesting about those plugins is that vendors, companies, and open-source projects actually allow for developing the plugins for you. You just need to pick them, grab them, add them, configure them—kind of an easy job using the new plugin system. And based on that, you can get the plugins into your system. We can see the research by Roi, and you can see how many companies use how many plugins. You can see the on-page PDD, Tech Radar—all those plugins are very popular.

Now we can see a trend very recently that all the vendors in the industry are running after creating more and more plugins into their own system. This is kind of amazing because it allows us to create this single pane of glass. So, we have the Backstage plugin website where we can see all the plugins, and we can integrate everything from CD systems to observability systems to funny Dev quotes that we can add in order to get some funny Dev time when we get into the portal. Also, if you want to add an event check into the portal and add gen AI-powered chat, everything can be done, and there are really a lot of plugins so you can use them. That’s super powerful, and based on that, you can build your own portal, creating everything you like on this canvas. You can see that, for example, we can run the Cubes SP plugin to get security or to get a reliability score from Komodor, DataDog, Argo CD, the relations, and about created initially like originally by Backstage—everything in a single pane of glass. And any one of you can basically build what you want on your own.

Now it’s a good time to summarize just what we saw. So, we talk about the single pane of glass, and now we can want and embed more and more plugins. And we have a few minutes for questions. So, are there any questions from the audience? Oh, over there, there is a mic pretty close to you, so there stands with mics. Oh okay, thank you very much.

Good morning. I have a question regarding the plugins because I’m a newbie to Backstage, and I have the feeling it’s always you need to build the complete image every time you need to configure a new plugin. Do you plan on the roadmap to make it more dynamic, maybe? Thank you. Yes, this is what the new plugin system is all about—to take the static configuration or the static way that you had to add the plugins, build the image, and then deploy it, and move it to a way where you can deploy one image and then add the plugins after. That will make it more similar to our plugins used in other systems that you know.

Any other questions? Oh, over there. Maybe this is just a point of clarification, but some of these cards on the screen here, are all part of Open Source Backstage, or does this also include some of the paid plugins? So, some of them, for example, the Cubes SP, are open source or will be open-sourced. It’s still very much in an alpha state, but we’ll get there. Okay, and I was specifically looking at like the Cube Scape, Argo CD, and DataDog cards. So, Argo CD is fully open source if it’s an open-source project, so the plugin is always open source. In most cases, the plugins are open source, but sometimes you need to add it to a paid product, so you need to pay for the product, and the plugin is free. Almost all of the companies in the market are investing in those plugins, so if there is no plugin available, we see a big trend of people trying to add it. Sometimes it’s the company itself, and sometimes it’s the open-source initiatives just for the plugin.

And I just want to reinforce what Guy said about that. Many of these are open-source projects, so if you need a plugin, we can build it within the community. You don’t have to wait for it to come out. I think I got confused about when the scorecard was brought up as well as transitioning into these other

plugins. Thanks. So, you can use scoring and quality in different ways. Scorecard is one of them, and those plugins that are more on the overview page offer different ways of doing that. You can decide whatever you like. This is the nice thing about it—you can integrate one plugin, take one component from it, and then integrate another plugin because you want capabilities from the other one. And what’s really cool about Backstage is this mix-and-match approach, so you can mix and match whatever you like in order to get to your desired state at the end.

Okay, nice. I actually have kind of two questions. The one that kind of woke up my curiosity was those gen AI environments. Can you maybe elaborate a little bit more about them because I honestly did not get what the template is capable of building in the end. And the second thing is the usage of these cards and Soundchecks. Can we already use some kind of metrics besides it, especially besides these scorecards? Don’t any kind of metrics already provide the usage, users this and that? So basically, those are two questions that I have.

So, I will start with the Gen example. Gen AI is just an example of how we can deliver high-quality application templates into the organization even if some people in the organization are not subject experts on that matter. For example, creating Gen AI applications is something that we all want to do, but it’s very hard. So what AWS did do, they just created a template, something that you can plug into your Backstage, use the plugins, and then spin up Gen AI applications. So what it essentially does is it will create the AWS resources. This isn’t that example but it can be whatever cloud provider, whatever Kubernetes provider you want. It’s just an example of creating all the resources you need in AWS to spin up an OpenAI or AI application connected to a database. Is it clear?

Sorry, so wait, um, if I understood correctly, does that also allow me in some kind of way to even build on top of it, my let’s say clusters or something through this AI? So you can take the template to do whatever you like. This is just one example, but there are other examples out there and you can always create your templates based on what you want to get, obviously, on your cluster or whatever cluster you like.

Okay, okay, perfect. Okay, yeah. Uh, we’re at a time, so thank you for the insightful questions. You can talk to Guy and me after and throughout the conference. We will be hanging around the Komodor Booth, the ARMO booth, and the Kubescape kiosk, so come find us and talk to us.

But let’s just kind of close this. Land This Plane—wrong metaphor for Backstage, okay, sorry. Think of Backstage as a canvas, and you can build and enrich it with plugins that suit your needs and follow the example of these top-performing companies that are using Backstage to reduce cognitive load, to increase quality, and speed up deployments.

Thank you so much for coming this morning!