It's All About Observability: Jaeger, OpenSearch, and OpenTelemetry



Recovered from

In this podcast episode, Katherine Druckman interviews Jonah Kowall of Aiven about observability and his involvement in projects like Jaeger, OpenTelemetry, and OpenSearch. They discuss the importance of project sustainability and the challenges of attracting and retaining contributors. They also touch on licensing controversies and the evolving ecosystem of open source projects. Kowall emphasizes the need for collaboration and communication within the open source community and encourages individuals to get involved by joining Slack channels, contributing to GitHub repositories, and attending events like KubeCon.

"I've really seen this ecosystem change a lot over the years, and there's always cycles of changes, and I think we're about to go through a pretty big cycle of change, and there's a few different things that are causing that. One is that there are new technologies coming into play. With OpenTelemetry, the data collection is all normalized. So now as a startup, I don't have to care about all of the data collection. So I think we're going to see a big change in the next few years of new companies coming up that really take that to the next level."  — Jonah Kowall


Katherine Druckman: At the recent KubeCon in Paris, I sat down with Jonah Kowall of Aiven to chat about observability and why it's such a hot topic, and his involvement in projects like Jaeger and OpenSearch. We also touched on OpenTelemetry and covered licensing controversies and challenges with project sustainability. It's a good one.

Hey Jonah, thank you for carving out a little time to talk to me today. I appreciate it. Can you tell me a little bit about who you are, your background and what you do, and what you're doing here at KubeCon?

Jonah Kowall: Sure, definitely. Thanks for having me on. I appreciate it, and I love your little setup and the things that you're doing here at Intel.

Katherine Druckman: Thanks.

Jonah Kowall: Thanks for that. A little bit on my background is I come from a background generally of observability. I was previously an analyst at Gartner covering that. I've worked at several vendors, and I've always been a big fan of figuring out how things work. So, observability has been near and dear to my heart for many decades and a lot of time. My day job today at Aiven is I run product management, design, docs, and some other things at the company, but my fun job is that I like to work in observability. So, I specifically am a maintainer of the CNCF project Jaeger, and I'm also part of the leadership committee and a regular contributor to the OpenSearch project. And then I dabble in OpenTelemetry because it's related to Jaeger, but it's not something that I spend a lot of time on these days just because I don't have enough. Otherwise, I would.

Deep Dive into Jaeger: The Observability Tool

Katherine Druckman: Right. None of us do. That's very cool. Well, okay, for those who don't know, let's start with Jaeger. Tell us a little bit about the project.

Jonah Kowall: Yeah, so Jaeger is one of the oldest CNCF projects and it was donated by Uber to the CNCF and has been graduated for several years. So, it's mature, well adopted. It is the only tracing visualization and data storage project that is truly in an open source software foundation today, which is somewhat concerning because it's a very important signal in observability, but I think that that is going to change in the future as there are some new projects that will hopefully be part of foundations in the future.

Jaeger's really key to visualizing and understanding traces, and anyone that gets started in tracing and open source uses Jaeger. Sometimes they stay with Jaeger, sometimes they go commercial eventually. There's lots of options when you start working with OpenTelemetry.

Exploring OpenSearch and Its Ecosystem

Katherine Druckman: Very cool. And then tell us a little bit about OpenSearch.

Jonah Kowall: Yeah, so OpenSearch was built due to the licensing change of ElasticSearch a few years ago, and AWS was a big part of creating and has been the main contributor to OpenSearch today. However, there are a lot of different companies now involved in building it and using it, and we have a great ecosystem with other cloud providers and other constituents contributing significantly. At Aiven, we have full-time people in our open source program office working on it, because it's a key part of our portfolio. You have companies like Oracle Cloud who have big contributors and care a lot about it, and lots of other organizations. We're in the process of trying to move it to a software foundation, which is something that Amazon has not gone through. So it's a bit of a learning experience, and we'll get it there, and I think it's going to continue to be a great viable technology.

The Impact of Licensing Changes on Open Source

Katherine Druckman: Yeah, it's funny, I feel like a few years ago there was a major upheaval in terms of licensing and there are obviously a few things happening more recently, but I feel like it was a major, major point of discussion several years ago, and a lot of interesting things came out of that. What do you think about the licensing controversy and how it shapes the ecosystem?

Jonah Kowall: So yesterday morning when I woke up, one of the projects that Aiven offers as a managed service is Redis. And I woke up yesterday and found out that Redis has changed their license.

Katherine Druckman: Again?

Jonah Kowall: Yeah. Now it's officially not open source. They've moved it entirely proprietary, and this is brand new news, hot off the presses. So I've been having to do a little damage control because our customers are asking, and people inside our company are asking, and we're trying to figure out what we're going to do.

Katherine Druckman: That's very interesting. I followed that conversation quite a bit back in 2018, 2019 when it was happening, when they changed to the server side public license I believe was the first one, and then it went through another. There was something Redis Source Available. I can't remember the name of the license.

Jonah Kowall: But the core was still open source.

Katherine Druckman: Yes, yes. And then they had a-

Jonah Kowall: Now the core is no longer open source. That's the new change.

Katherine Druckman: Right. Yeah. That's very interesting development. I didn't know that. Thank you.

Jonah Kowall: It's brand new, and at Aiven, everything we do is open source based in terms of our product offerings. Including things that we build, we open source it. Some of our competitors even use the things that we open source, which is totally fine.

Katherine Druckman: Yeah, sure. That's the nature.

Jonah Kowall: But this week we launched actually our first non-open source project. We took a technology called Dragonfly. They're here at the KubeCon.

Katherine Druckman: Yeah, I know Dragonfly. Yep.

Jonah Kowall: And now we offer Dragonfly on Aiven, which we, from them, it is not fully open source, but it's source available and that's Redis compatible. So thankfully we do have a path forward, but we-

Katherine Druckman: No, that's interesting. Okay.

Jonah Kowall: We did that for performance reasons because of Redis' scalability, but now we have a path forward potentially with the rest of our Redis customers as well.

Katherine Druckman: That is interesting because again, it's always a risk, right? There's a risk inherent in the kind of work that you do where you depend on a project and yeah, it'll be interesting to see what the community reaction is to the change.

Jonah Kowall: Twitter has been interesting to take a look-

Katherine Druckman: Yeah, I hope because I've been here in the fishbowl and I've not followed, I'm going to have to look this up.

Jonah Kowall: Yep.

Katherine Druckman: It's fascinating.

Jonah Kowall: And that's why the CNCF is so important because all of these projects like Jaeger and OpenTelemetry and Prometheus and everything around Kubernetes, they're protected from someone hijacking the project.

Katherine Druckman: Sure. Yeah, yeah.

Jonah Kowall: However, the problem that we have, and I can tell you this from Jaeger too, is that it's hard for us to get maintainers and consistent contributors. We have four maintainers of Jaeger, and I would say that half of the team doesn't spend that much time on it. And as we age and get busy, we need new blood coming in.

The Challenge of Sustaining Open Source Projects

Katherine Druckman: Sure. Project sustainability... Licensing, very important, very important conversation. Equally important, project sustainability and project health and mentoring your replacement. Right?

Jonah Kowall: Yep. So we participate in both the Linux Foundation Mentorship program, which is currently running. We're there in session and we also participate in Google Summer of Code where they actually pay interns.

Katherine Druckman: Great, that's fantastic.

Jonah Kowall: The problem is that when those internships end, they tend not to continue contributing. We want them to become contributors, hopefully maintainers, but it hasn't happened, and we've been doing this now for three or four years. The internships, they help with us getting the project moved forward. We work on bigger initiatives and projects that they help a lot with. However, we don't get new maintainers or ongoing contributions, which is, it's a big challenge. I'd love to figure out how to solve it.

Fostering New Contributors and Community Engagement

Katherine Druckman: Yeah, I would love your thoughts on... I'm sure you've given it a lot of thought. Do you have any paths of exploration that you're thinking in terms of how do you get people hooked in? It's a tough conversation. How did you get hooked in? Why are you the open source enthusiast that you are today?

Jonah Kowall: I think it just became something that I felt was valuable to the community, but I initially started because of work. My previous startup, we were using Jaeger and built it into our product. So I got involved in it because we actually contributed a whole bunch of new features to Jaeger. That's how it started. And I stuck with it even when I changed jobs, because I like the people. I like the community. I love coming to KubeCon. I like speaking. We have a ContribFest on Jaeger coming up in about an hour, so I'm going to go there.

Katherine Druckman: Oh, okay. Cool.

Jonah Kowall: We try to get more people. I've had discussions with a few folks that do want to contribute to the project here at KubeCon. So it's a way for us to get interest and people on board. We're trying everything that we can to solve that problem. However, I don't have the answer.

Katherine Druckman: Yeah, it's tough. I think I might be a bit of a weirdo because I came into it not for work. I came into it because I had a personal project that I wanted to work on, and that's how I got hooked into using Drupal. And then I started working for Linux Journal, and it was always a thing. Anything I did that I would consider a contribution was more of a thank you because I feel like all these people taught me a lot of things, but it's so hard to define. It's hard to define what keeps you coming back. Somebody once called me a culture junkie. I don't know if we need to just find all the culture junkies in the world and reward them with the great open source culture.

So speaking of encouraging new contributors, maybe this is a great opportunity. If you could share what types of contributions could you really use? What types of people out there might be a great fit for joining the community?

Jonah Kowall: So when we look at the maintainers and the people that work on it, we all tend to be more back-end people. We look a lot at the back-end and the way that the project works from an infrastructure and scalability perspective; that's where our expertise is collectively. We have no one that's a UI person, that's a front-end person. We get great contributions and one-offs from people that want to add something to the UI. But anyone with front-end experience, UI experience, the Jaeger UI is something that people really love when they see it.

It has really nice visualization. It's very clear. It's often better than many of the commercial tools. And that appeals to people because when you're dealing with highly complex, abstract ideas like tracing, if you visualize it clearly, people get it and they say, "This will solve my problem, and I understand how to see the issue," because it's a very visual view of what's happening inside your software. And so that is the thing that makes Jaeger unique, and it's the thing that we don't have contributions to today.

Katherine Druckman: That's such an important topic too, the front end and the user experience and all of that, it's maybe an underserved area, I think, in the open source community. I would love to see people get more involved. Obviously, there are people doing great work out there in various projects, but well, we could all use those more, right?

Jonah Kowall: Yeah.

Katherine Druckman: If nobody's using your project because they have bad usability, then that's a problem.

Jonah Kowall: Yeah. And a lot of the tools that you see in the CNCF don't even really have any UI or anything at all. In OpenSearch, there's a lot of work that goes on to OpenSearch Dashboards, which is the UI for it, and there's a lot of things that you can do within it. So in many ways, we work in Jaeger, we're compatible with OpenSearch and we use it as a database. And they also have built things in OpenSearch Dashboards to visualize Jaeger data. So we say, people can use OpenSearch Dashboards, they can use Jaeger on the same data in the same schema. In many ways, it's interoperability between different open source and giving people choice. That's what open source is about.

Observability Trends and the Future of Open Source

Katherine Druckman: Yeah, absolutely. I agree. So it's something that I've noticed at this event. We've mentioned we're at KubeCon, I've mentioned the fishbowl, because I'm in here, we're here at the Intel booth and our little podcasting studio, which is great fun. But this event, the theme for me and the people I've talked to has been observability, a little bit of cost savings and that kind of thing, but definitely observability. Whereas I think in Chicago, I would say the theme was AI. Everybody wanted to talk about AI, but this time everybody's talking to me about observability. Why do you think that is?

Jonah Kowall: I've really seen this ecosystem change a lot over the years, and there's always cycles of changes, and I think we're about to go through a pretty big cycle of change, and there's a few different things that are causing that. One is that there's new technologies coming into play. So good examples of that are profiling, which is becoming part of OpenTelemetry, which adds a signal that in commercial application performance monitoring tools, profiling has always been part of them.

Katherine Druckman: Sure.

Jonah Kowall: Going back over 10 years. In open source, profiling has always been a separate thing. It was never integrated well. So I think that's the big piece that's starting to change is you're going to have profiling part of OpenTelemetry, and then databases are changing, and that's going to allow companies to build more cost-effective and scalable solutions around this technology. With OpenTelemetry, the data collection is all normalized. Now as a startup, I don't have to care about all of the data collection, which I can tell you from my experience at AppDynamics, we would spend a huge percentage of our engineering budget on data collection. Now I could just focus on visualization, data storage, and optimizing that. There's some great new startups that are doing just that, and they're really focusing on that usability and getting value out of the data instead of collecting it. So I think we're going to see a big change in the next few years of new companies coming up that really take that to the next level.

Katherine Druckman: Yeah. Well, to tie it in with the previous hype cycle, AI is all about, just at the end of the day, processing data quickly, and in a way that's maybe more usable or more natural, and what good is all the data if we don't know what to do with it?

Jonah Kowall: Yeah.

Katherine Druckman: So that makes a lot of sense. I wonder if you could tell me a little bit more about OpenTelemetry.

Jonah Kowall: Yeah, so OpenTelemetry is that normalized data layer that allows people to collect the three main observability signals, logs, metrics, and traces. In Jaeger, we're primarily focused on tracing. We do a little bit with metrics. We don't unify them. When you look at commercial tools, they try to unify all of those signals. And OpenTelemetry is that unified data collection for all of that data. But it's evolving, there's work going on with adding real user monitoring into OpenTelemetry. That's been going on for a while. It's starting to move faster. There's the work on profiling that's starting to accelerate. There's been some nice contributions from vendors to get profiling into OpenTelemetry. There's a new event spec that deals with events that are slightly different than logs.

So the project continues to evolve, but continues to also get more complex. People come to us and say, "Why is OpenTelemetry so complex?" It's like, "Well, it deals with a lot of different data and a lot of different signals that are used for different things, and it's an inherently a complex problem to solve." So OpenTelemetry just tries to unify that, and there's a lot of great tools that try to make it easier to manage, but it's a complex ecosystem right now.

Katherine Druckman: Yeah. Yes. That is an understatement. Yeah, I talk a lot about the complexity of this great ecosystem of which we are part, right? It's a lot of moving parts and it's hard to keep track of. So speaking of that, I feel like there probably is a bit of competition. We talked about project sustainability and recruiting new contributors. There must be quite a bit of competition given the cloud-native landscape, open source landscape, if you want to call it a single body. There is a lot of competition for contributors. You have to get contributors excited about solving the problem that you're trying to solve together. I wondered if the projects that you're involved in have a bit of overlap, inner communication, do you see people contributing to all of them, or are you just an overachiever?

Jonah Kowall: Well, I don't spend much time on OpenTelemetry anymore as I used to, just because I don't have time to do all of it. It's still relevant to Jaeger, so I pay some attention, but I'm by no means really part of the project anymore. I have a lot of friends that are, because we have a lot of people that work across different areas. Similarly, on the metric side, I work and talk a lot with the Prometheus and the Thanos community there, and there's a bit of overlap and competition in some of the projects. So a good example is Fluentd and Fluent Bit are very popular for logging, but OpenTelemetry does logging. And Fluentd and Fluent Bit are now doing some metrics and a little bit of tracing, and so they have some overlap. But when you talk to a user and you say, "Are you doing tracing?" The answer is OpenTelemetry. If you only want logs, the answer is probably Fluentd and Fluent Bit.

So although there's competition, there's use cases for things. We also have other problems where Prometheus is the de facto way that people deal with metrics, and they don't always agree with the OpenTelemetry metrics people. And there's been a little bit of fighting between the two sides, and they sometimes work well together and sometimes don't. And that animosity is just not good to see because everyone just wants things to work well, and we do want to adopt things that are well established in the community, but that's not always the opinion of everyone.

Enhancing Collaboration in the Open Source Ecosystem

Katherine Druckman: Yeah, because Fluentd being good for logging, and... So we see a lot more specialization. When you're solving a very specific problem and you're in it, you become a bit blind to what's around you. And I think a lot of these projects could benefit from a little bit more communication within the ecosystem and a little more cross-pollination maybe, if that makes sense.

Jonah Kowall: Yep.

Katherine Druckman: I wondered if you have any thoughts on how that might be achievable. Again, when you have competing interests, and you have... But the spirit of open source is that we all are in this together and we do share the burden. So there are common goals, and I wondered if you can see a way that that might move forward more efficiently.

Jonah Kowall: Well, that's why KubeCon's so important because we have an observability day before, and we sit next to each other in our little kiosks in the CNCF area, and we talk to each other, and I talk to my friends that work on different things, and we talk about problems and we talk about challenges, and getting everyone together is actually the best way to collaborate. Even though no one, and I don't work in an office, and a lot of people don't want to, getting people together helps you brainstorm and think about new things and try to solve problems. And that's why KubeCon is such a great place to do that because there are 13,000 people here that care about open source, and maintainers of all of these projects are here, and we need to talk to each other more so that we don't have our blinders on, and we don't just look at one problem and we think about the entire ecosystem.

Katherine Druckman: Yeah, I think that's really valuable.

Jonah Kowall: Yeah.

Katherine Druckman: So one, I want to ask you if there's anything that we didn't cover that you wanted to.

Jonah Kowall: No, I think we touched on a lot of key points. I was lucky enough to spend some time with Kelsey Hightower a few weeks ago, and we talked a lot about this open source sustainability, and how all of the companies that consume open source need to give back to the community and need to be part of this.

Katherine Druckman: Yes, 1000%.

Jonah Kowall: But I don't think that that is going to really make it happen.

Final Thoughts and Advice for Aspiring Contributors

Katherine Druckman: Right, yeah, it's doing the work, putting in the work, for sure. So as a final thought, I wondered, what advice would you have for people that would like to get involved in some of these projects?

Jonah Kowall: So we all are on the CNCF Slack, and we all tag things like Good First Issue in our GitHub, and there are always ideas going on that anyone can pick up and work on. We have a Contrib Fest here at KubeCon. We run tutorial sessions, we do maintainer talks, we do whatever we can to get the community engaged. Just being at the Jaeger booth, having people come by that have ideas or that want to contribute, that's the way that we try to reach out to the community. But the Slack is a great place to come to, to ask questions, or just open an issue on our GitHub, even if it's "I have this idea, does it make sense?" And we can talk about it there, or we can talk about it on Slack. That's how we get a lot of the contributions that are really useful.

Katherine Druckman: Yeah. Well, fantastic. Thank you so much. This has been great. I've learned a lot just in, gosh, not so many minutes, so I really appreciate it, and thank you very much. And I'm pretty sure the listeners do too.

Jonah Kowall: Thank you for having me, and I hope we get to do this again.

Katherine Druckman: Awesome.

You've been listening to Open At Intel. Be sure to check out more from the Open At Intel podcast at and @OpenAtIntel on Twitter. We hope you join us again next time to geek out about open source.

About the Author

Katherine Druckman, Open Source Evangelist, Intel 

Katherine Druckman, an Intel open source evangelist, hosts the podcasts Open at Intel, Reality 2.0, and FLOSS Weekly. A security and privacy advocate, software engineer, and former digital director of Linux Journal, she’s a longtime champion of open source and open standards.