Danielle Cook, VP of Marketing, Fairwinds: Welcome. I am here today with Rishi from Clover Network, Tristan from PagerDuty and Kendall from Fairwinds. And we're here to talk about Kubernetes, PagerDuty, Fairwinds and how we are all working together. Thank you all for joining me. I first want to start off by asking you all to tell me a little bit about your role and what you do with Kubernetes and your organization.
Tristan, do you want to kick it off?
Tristan Bates, Engineering Manager, SRE Platform, PagerDuty: Sure. I have an engineering manager for the SRE platform engineering team at PagerDuty. The SRE platform team is responsible for the implementation of Kubernetes. We deliver Kubernetes-based platforms so that our engineering teams can focus on building features without worrying about all the complexities of the underlying infrastructure.
Rishi Malik, VP of Platform Engineering, Clover Network: Thanks. I run platform engineering at Clover Network. We are a point-of-sale company that makes it easy for small to medium businesses to run their business. We do a lot with Kubernetes. It’s really the foundation for our next generation platform. Similar to what Tristan said, it's a way to equip our application teams to build and run quickly, get rid of a lot of complexity that they had to deal with traditionally. Awesome.
Kendall Miller, President, Fairwinds: I'm Kendall from Fairwinds. Fairwinds has been in the Kubernetes space for 7 years, that can't be quite that long 6 and a half years, building and maintaining infrastructure for lots and lots of people for a long time. Today we develop software to put guardrails around that. As Kubernetes becomes more accessible, standing it up is no longer the hard part. It seems like everybody can do that. It's once it's stood up that everybody finds they're in trouble. Trying to put guardrails around it. Best practices for security, efficiency and reliability is where Fairwinds plays.
Cook: Awesome. So now I just want to ask a few questions and get anyone to jump in and answer. Let's have a discussion. The first thing is around what are the kind of things you need to monitor in Kubernetes and what do you need to be alerted on?
Malik: For us, we monitor everything. We get alerted on quite a bit, but more specifically we have three different aspects to what we care about: you have those traditional infrastructure side of monitoring which our platform teams are mostly getting alerted on.
We have the application side which is specific to what each individual team is doing on the platform. And what makes sense for them to be alerted on based on their customer experience.
The third piece is actually what we do with Fairwinds. And that's really the scale side of things. We have some of the application performance type of alerting we get through Fairwinds. But one of the big things we actually care about is cost. A big part of what we've done by moving to Kubernetes, by integrating Fairwinds, by tying that into our PagerDuty setup, is really geared around self-service for all of our application teams. That’s great for the engineering side because they're quick, they can run fast, there is no bottleneck to the platform teams.
Fundamentally, it's a shift in how we do business. For us to be able to say this application team, this service, this product, this product feature has a specific cost and put alerting around that, helps us to know what we're doing as a business and where we need to focus and where we need to change what our investments are.
Bates: One of the big things we monitor, and that's important to solve at an organization level are security and compliance standards that need to be in place for every team. We don’t necessarily want the teams to know about it, we want to make it easy to do the right thing without blocking productivity. That's why we use Fairwinds Insights, its Admission Controller and reporting to manage Kubernetes security and compliance at PagerDuty. We even integrate it through the entire lifecycle to alert teams if they are doing something that's out of what we're supposed to be doing.
Miller: I want to add Tristan, thank you for doing that. As a PagerDuty customer, I'm glad to know that you all are keeping that lock down and are being tuned into things like security alerts, etc. That's a lot of what we spend time on too is like where are there deprecated APIs, where are there out of date add-on, to make sure we are staying up to date. So much of security problems come from just running an outdated version of something that you didn't even know is there. And so that's a big part of it. I'm really pleased to hear both of those things from both of you and a thank you Tristan for making your product secure as the person who relies heavily on it.
Cook: That’s important: Fairwinds is a customer of PagerDuty, PagerDuty is a customer of Fairwinds and then we have Rishi from Clover who works with both PagerDuty and Fairwinds. So it's a tight group of people here able to comment on the different technologies and how they're helping everyone to do business. My next question, and this is for you Rishi, what are the features you really like in Fairwinds Insight and PagerDuty and how they work together?
Malik: Sure. For me, it's all about that concept of self-service and equipping application teams to be able to till their own destiny. With Fairwinds specifically, we get cost analysis, security alerts and honestly, one of the biggest things we get is the pane of glass, the dashboard, that’s easy for our applications teams to login and see what’s going on. To login and see what’s there versus using some more complex tools that platform teams are typically more familiar with. That ease of use, reducing friction is absolutely key for us because it’s a lot to put on applications teams that don’t necessarily have the same level of experience as the platform side. Fairwinds makes it easy and accessible.
It's similar on the PagerDuty side. We get the value of the basic alerting and paging. The biggest thing where we see large value is stuff like deduplication, where a team may have hundreds of servers and we see something happen at an aggregate level – a cloud provider has an issue with some aspect of their service – that sets off alerts on every single instance. The deduplication is huge as it’s something we had to manage in the past.
It took time and effort, but we now get that with our PagerDuty integration. We also again, get more of that self-service aspect where teams can define their escalation policies. Our platform team has one that is super validated and tight and application teams are also able to set what makes sense for them and their customers. For that stuff that isn’t as critical, we can escalate differently. We get the visibility to be able to say everyone has an escalation policy, everyone has a bare minimum escalation policy and then equip the team with whatever else they need on top. It really comes back to that self-service idea. Both these tools [Fairwinds Insights and PagerDuty] help us drive that across the organization.
Cook: That's great. And I love that it echoes what Tristan said – you are putting guardrails in place. Your developers don't even need to know the details of the guardrails. People are self-servicing and able to get the job done faster because of Fairwinds Insights and PagerDuty.
Bates: Yeah. We practice a whole service ownership model here, but we realized that doesn't mean you have to understand everything - you just need to shape and own your thing. Using tools like Fairwinds Insights and of course, PagerDuty, makes all that simpler so users can focus on the important things.
Miller: I would add to that – service ownership in a modern DevOps world, felt like a pipe dream until things like Heroku came along. People don’t understand all the bits happening underneath Heroku – they don’t need to – Heroku is the tool that enables a developer to push a button and ship through to production. Nowadays, everybody's out building their own version of that increasingly with Kubernetes. But you don’t want to have to educate all of your developers on all the things Kubernetes. Instead, you want to be able to put bumpers up on the bowling lane so they don't have to be very good at avoiding the gutters, they can just throw the ball down without issue.
Cook: And I think the point there is that you could have your developers spending time becoming experts and updates, but it takes away from the value you're providing from your product in your application. It’s that balance of we need you to do the right things with this, but we don't need every one of you to be an expert on Kubernetes because it’s complex.
Malik: What I love about this shift – you can call it shift-left, call it self-service – whatever it may be – It’s the fact that we are bringing developer empathy to the forefront. Kendall, you mentioned Heroku and yes, they made it simple and easy, but fundamentally, what they did was they made it friendly and nice. It was the concern of the developer, the developer teams’ experience to ship code and that was the driving factor.
As things have got more complex, as cloud migrations have happened, I feel like we've lost that a little bit. We're trying to get back to it with equipping teams to run faster because that is how most organizations have figured out how to scale. You want to be able to hire engineers working on the revenue side of what you're doing.
But I like the fact, I love that, we can have this conversation and bring in tools and talk about how we can make life better for developers. By extension that makes life better for us as a business. It really is that ease of use, that empathy. You are under pressure as an engineer to deliver products and features, and to ship code. Let's not just make it super hard, complex or as old school as it used to be. Let’s make it modern, and make it easy to remember that you're driving a lot of value for the business. Let's do what we can to help you out. I love that aspect.
Bates: To add to that, I just want to say how nicely Fairwinds fit into that mindset. We looked at what it would have taken us to build all these features ourselves and glue together various tools, products and open source, glue code. We would have spent our entire budget and roadmap for this project running Kubernetes and future infrastructure on just doing that versus plugging in Fairwinds Insights
And being able to take care of that without too much effort.
Miller: That’s a sound byte I can take and ship right there, Tristan.
I think what's interesting about it, I made the comment thanking Tristan for making your product secure but it occurs to me too that I’m often using a Clover point of sale so having that be secure as well, if anything, probably has more direct effect on me in some ways.
To the point about shift left - at any kind of size and scale, it’s just simply impossible for a platform team to maintain security or any kind of standard alone. When you have tools that push that left so that the service owners are actually aware of, stopped by, guard railed with, kept up to date, whatever all the words are so that things get shipped through and it's not just this core group of platform people or DevOps, or SREs, or whatever the organization is calling them, that’s a big deal and it really is impossible at scale.
Cook: We’ve hit a lot of good things today around making the developers ship applications faster, arming them with the tools they need to be compliant, be secure, use Kubernetes. All of those things are really great and it's the technology that PagerDuty is providing as well as Insights together that is really making a difference in the organization. Thank you all for joining.