Dreamforce 2024 | Bob Wise | Heroku

Clips
News
More from Dreamforce 2024

Bob Wise

CEO

Heroku

Heroku’s role in Salesforce’s ecosystem: Enhancing developer productivity and simplifying deployment

Enterprise tech is built on the back of software developers, and Heroku is Salesforce Inc.’s testament to developer empowerment. Its integration into the larger Salesforce platform seeks to bridge operational gaps while enhancing app deployments and developer productivity.Heroku’s Bob Wise discusses the platform’s value for developers.“If you have some code that you’ve written in any language and you want to run that code in production, there is a lot of work to be done,” said Bob Wise (pictured), chief executive officer of Heroku, a Salesforce company. “What you get in a lot of those cases is a bag of Lego bricks

play_circle_outline Heroku simplifies deployment process by handling authentication, storage, load balancing, etc.

play_circle_outline Heroku integrates with Salesforce for pro-code development and sophisticated customer needs.

play_circle_outline Heroku incorporates Kubernetes under the hood to provide standardization and adaptability.

play_circle_outline Salesforce is a major user of Heroku, running apps like Trailhead and BT org.

Info

Bob Wise | Heroku

Bob Wise

CEO Heroku

search

>> Hello and welcome back to Dreamforce 2024, live from the NYSE studio that we built out right across from Moscone. Really bringing all of the news that's going on in the happenings from Salesforce Dreamforce 2024. And we're digging in deep. And I think this is one of the things that's near and dear to my heart, being that DevOps developer, digging into containers and Kubernetes. I think one of the things that you always find interesting is some of the companies that are owned by Salesforce, and one of them is Heroku. And we're joined by Bob Weiss, who's the CEO of Heroku, and really get to dig in here about what's going on and how you're really helping bridge that developer gap to more of a container Kubernetes type of offering that.

>> Right.

>> Yeah, so welcome on board.

>> Thanks for having me.

>> So, help people understand Heroku that may not be familiar with how it fits in and how it fits into the whole family.

>> Certainly. Well, let's start with what Heroku does. If you have some code that you've written in any language and you want to run that code in production, there is a lot of work to be done, if you start with building an AWS, getting an EC2 instance and wearing that authentication and then storage and all these bits. So what you get in a lot of those cases is, you get a bag of Lego bricks. And what you have to do before you can even deploy your application, is go build a whole model. And what Heroku does is assemble all of those things into a single piece. So, we can connect to your GitHub repo, and then you can take your code and just push it to production, and we take care of all the hard stuff.

>> So you're broke-

>> Scaling, load balancers-

>> Yeah, all the clustering-...

>> , all that stuff....

>> and all of that.

>> Correct.

>> Yeah, and I think that to me is really interesting, because I think that both of us spent time over at AWS and have built stuff there. And again, always building that layer cake. And you always had to build up to get to your platform that you were building. And then you got your platform, but then you had to build the apps on top of that. So, there was a lot of engineering that went into that. Where does this kind of join in with what else is going on at Salesforce as well?

>> So Heroku was acquired a long time ago, 14 years ago, well before my time here at Salesforce. And the original idea behind the acquisition is the reason it's important to Salesforce today. Which is, although Salesforce broadly is very dedicated to no code, low code, easy configuration, all of those things, there's always going to be that time when you need to go and write a serious amount of code. Or maybe you have an existing system that's already built and you need to integrate it into the rest of Salesforce. And so having a PaaS, a platform like Heroku that is very oriented towards what we call pro-code, the professional developer people who write code every day, to kind of extend and add all the superpowers to Salesforce is really important. So, we're an important part of the portfolio for particularly the more sophisticated customers who need to write a lot of code in and around Salesforce. But we also have a lot of customers who use Heroku as their general purpose platform.

>> Yeah. No, I think to me, when I looked at it and I looked at the whole capabilities from the various different languages, it's not like, "Hey, it's one language. It's just Java or JavaScript or .net." It's like Ruby, it's like Go, it's like everything, it seems like that.

>> Yes. Right. We want to meet the developers where they are. And I think you know, you've been in this space, developers love the language and the frameworks that they love. And so, we try to support all of those.

>> And having a previous startup of mine, we actually built on force.com and actually used the platform, so I know how flexible it can be. And help us understand how it ties into kind of that Kubernetes environment, because it has that all in there already wrapped in.

>> It does. So, Heroku was the original cloud native. It was cloud native before cloud native even became a thing. And you might remember when originally the cloud native computing foundation was going to be called the Kubernetes Foundation. And we all decided it really ought to be something more general purpose. And so, there's been a lot of standardization around Kubernetes as kind of the underlying distributed operating system. But Heroku had built a proprietary, a really awesome proprietary version of that. And so, now we're at this point where we are moving to run Kubernetes under the covers of everything Heroku does. And so, we're providing standardization on top of that. And that's giving us the ability to move a lot faster and adapt to a lot of standards that customers really want, things like OTel. This is all kind of deep in the DevOps platform engineering-

>> Absolutely....

>> world. But it allows us for the customers, especially the customers that are using Heroku as their overall platform, those are the things that they really care about.

>> So, it'll help them tie into their observability and their platform engineering and things of that so they can standardize their processes across yourself. Because I have yet to meet one that uses one Kubernetes platform. Most have two or three. And I think they want to run it all in the same manner as well.

>> There's kind of a new wave of folks coming to Kubernetes and cloud native. The last 10 years, Kubernetes just had its 10th year anniversary, a lot of the folks there have been sort of the early adopters, the more sophisticated users. By the way, Salesforce is one of those. We have our internal platform called Hyperforce, which is a big Kubernetes, probably one of the world's largest Kubernetes deployments. But Salesforce has large numbers of people that work on that. And if you're not in this kind of tech-heavy segment or you're not interested in building a huge platform engineering team, hiring all these DevOps, platform engineers, Kubernetes experts, then you can come to Heroku and we'll help you with all of that.

>> So, it's the easy button to help people get up and running a lot quicker on that standardized platform. And I think, how does it really... Because a lot of talk and Mark is just offstage across the street, there's a lot of talk about agent force and AI. And how do you see that tying in with Heroku and what is all being announced this week?

>> Well, so it's sort of like what we were talking about before. If you're building agents, but then you have some of your own special sauce, proprietary code, things that you want to do to extend those agents, give them some superpowers, do something really interesting that requires you to write a lot of code, you want an easy way to do that. And Heroku was that way. And we also view that all applications are becoming AI applications. Now, that doesn't mean that every application is itself a model or an LLM. But every application has some need to connect to inference model, or to provide those kinds of services, whether it's RAG or access to some kind of easy inference model.

>> Yeah, it would seem, and I think again, you have some database services that can be deployed, you have all of the different pieces. So, people can come in and almost make their, what would be either a Helm chart or whatever, their configuration chart of their application and bring all those... And you must make that easy for people to go and do.

>> We do. One of the bits of Heroku magic is that we bring not just the compute pieces like Kubernetes does, but we also bring into all the database services, and Kafka and messaging. So if you're going to build an application, it isn't just building some code and deploying some code. You also need a database and you need a schema. And you need to securely connect the applications to the data. And you need the load balancing and all of those things. About half the Heroku business is running databases. It's a really big business. That's why we introduced pgvector support. If you're using RAG or some of these other techniques, you need to have a Vector database. And so we've been providing that now actually, for quite a while. So, that's another way that we've been supporting these kinds of AI superpower agents.

>> And can they bring their own database as well or their own models? It would seem like that.

>> You can. You can. And it's pretty common for folks, especially in Heroku, typically tend to be AWS customers as well. But maybe they do something with a GCP database service. So you can connect those things, and we make it easy to connect to additional services. How many database services does AWS have? Many, many, many.

>> It depends if you count Snowflake and Databricks and all of those on top of it.

>> On all of those. So, there's always a need to connect to those things externally. But we're also always looking for, where is the... We call it painting it purple, and that's our way to say, "Oh, it's part of the kit.", as opposed to being outside the kit. And so we're always looking for ways to extend that as well.

>> And I would assume being that it's a platform as a service, that they're using the databases that are painted purple already, it's super easy and all of that.

>> Super easy.

>> The backup and all of the different high availability is already built in for you.

>> It's all built in, it's super easy. The documentation is great, the support is great. So all of the pieces that you need together in the platform are what we do, and we do it really well.

>> And I would assume, like you said, you started talking about connecting up to other services like at AWS or something like that. Obviously, Salesforce has MuleSoft as well. And that must play some role in that.

>> We do. In fact, we have some really great customer talks going on this week. Lamborghini is here talking about using Data Cloud and Heroku together. Goosehead insurance is here talking about using MuleSoft and Heroku together.

>> Oh, great.

>> And so yes, if you want to use the... There's a lot of customer overlap between Heroku and MuleSoft, but not a lot of technical overlap. What MuleSoft does really well is not the same things that Heroku does really well. So if you need those two things together, that's an awesome match.

>> Yeah, just knowing Mule Soft, like if I want to connect up to a mainframe to bring data into an application that is on Heroku or being built out on Heroku-

>> Perfect. Perfect ....

>> it would seem like a very good way to bring that together 'cause it gives you that gateway. What else are you seeing? Again, you've been around Kubernetes for a while. How are you seeing the environment? Just the whole CNCF like you talked about, and not calling it Kubernetes Foundation, which I agree with. Because I think it's more about cloud native now, because like you said, 10 years on and with Kubernetes, it's solved by people like Heroku and others that really do it well.

>> I think the thing that I still see customers struggling with, and it's becoming even more important in this new AI era, is the ability to do, if you say CI/CD, can you continuously test the thing? And then can you continuously deploy that into production? That's really hard. Most people, I think, do a pretty good job on CI. That's pretty well understood. But the ability to build all of the pipelines that you need in order to make a change and get it into production without a human touching, it is a really hard chasm to cross. And a lot of companies haven't yet crossed it. That is one of the ways that having... So Heroku isn't just data database and compute services, it also includes the tooling that you need to run all of those pipelines. And the pipeline support is super critical. It's becoming even more critical now because if you think about agents building code, the effect on the team is very much like you just added a whole bunch of interns to your team, all writing code. All of this code needs to get to production. And if you haven't built the organizational muscle, you don't have the tooling in place to absorb all of those code changes that need to get to production, you're just going to be swamped in the soup of, "Oh, well, we have all this cool stuff, but how can we be sure that when we deploy it to production, it's still going to work?"

>> So, how does Heroku help with that CI/CD pipeline? I think you brought up testing for instance, and we were talking to one of the ecosystem partners about that just yesterday. And like you said, one of the critical things about AI is, how do you limit hallucinations? And part of that is the model training and other stuff that goes on in fine-tuning. But part of it is also making sure that the prompt engineering and all of that, which is all code at the end of the day. How is Heroku helping developers in general with that CI/CD?

>> So really the main point of CD, or I would say even CI, is to enable teams to go fast and not break things in production. But part of the go fast thing is, "Oh, I made a change and it would break something, but the tooling keeps it from getting into production." And so, that enables the teams to really go fast. And so, what's the effect of hallucination? The effect of hallucination should be, "Oh, this test case didn't pass.", which this is hard. This is very hard stuff to do, and this is the new frontier. But the final line of safety between the hallucination or any other mistake that a human makes or an agent makes, is having that great layer of testing in between. And so, you need a lot of tooling. If you're going to build a platform on top of Kubernetes, you actually have to go and figure out all of the pipeline tooling as well. And select from a vast number of different tools in that CNCF landscape, which there's over 200 projects now.

>> Yeah, it's unbelievable the projects and the mountain. There seems like there's a new one every day.

>> Exactly.

>> And which ones are in sandbox, which ones are actually being promoted, and things of that.

>> Each one of those things is sufficiently complicated that you could have a team of people that each one works on those things. And you still haven't integrated them all.

>> Right.

>> And so if you really want... So it's sort of the easy button in one way, but it's also the safe button in another. It's really important that when you pull all those things together, that you keep them operationally stable. One of those things changes, did it break everything else? There's a security vulnerability here. Did you even know that there was security vulnerability that needed to be patched? Heroku takes care of all of those things.

>> Right. And I think by saying what you said around Salesforce, one of the largest users of Kubernetes across with... The underlying, we actually were talking to David a little bit about that yesterday, around how that's actually built out on other clouds and things of that nature as a platform. It would seem like this ties in really nicely with the experience that Salesforce has running at scale as well.

>> Salesforce, if Salesforce were an external customer of Heroku, it would be our largest customer. And so, Salesforce uses Heroku in so many things. Trailhead, for example, runs entirely on Heroku, the BT org. That would be called the IT org, the CIO org in other companies, uses Heroku to deploy and run thousands and thousands of apps. I think it's over 10,000 apps now. So, Salesforce uses Heroku for this capability at scale.

>> As they say, I know it's not popular, but dog fooding it and understanding, and eating your own dog food kind of thing.

>> It's so important.

>> I think the better way to say this is sipping your own champagne.

>> Yes.

>> It's a better way of saying it. But that's always interesting. Well, hey Bob, I really appreciate you coming on. I think this really helps people understand just the expertise that Heroku and Salesforce bring to bear. Especially when people think of this, they can think of Kubernetes and they can think of Salesforce in the same terms.

>> That's great. Thanks so much. I'm happy to talk DevOps with you anytime.

>> Oh, we will. I'll take you up on that-

>> All right....

>> that's for sure. And thank you for watching this episode from Dreamforce 2024. We're live all day here at the NYSC. Stay tuned for more. And we'll be right back.

Bob Wise | Heroku

search

>> Hello and welcome back to Dreamforce 2024, live from the NYSE studio that we built out right across from Moscone. Really bringing all of the news that's going on in the happenings from Salesforce Dreamforce 2024. And we're digging in deep. And I think this is one of the things that's near and dear to my heart, being that DevOps developer, digging into containers and Kubernetes. I think one of the things that you always find interesting is some of the companies that are owned by Salesforce, and one of them is Heroku. And we're joined by Bob Weiss, who's the CEO of Heroku, and really get to dig in here about what's going on and how you're really helping bridge that developer gap to more of a container Kubernetes type of offering that.

>> Right.

>> Yeah, so welcome on board.

>> Thanks for having me.

>> So, help people understand Heroku that may not be familiar with how it fits in and how it fits into the whole family.

>> Certainly. Well, let's start with what Heroku does. If you have some code that you've written in any language and you want to run that code in production, there is a lot of work to be done, if you start with building an AWS, getting an EC2 instance and wearing that authentication and then storage and all these bits. So what you get in a lot of those cases is, you get a bag of Lego bricks. And what you have to do before you can even deploy your application, is go build a whole model. And what Heroku does is assemble all of those things into a single piece. So, we can connect to your GitHub repo, and then you can take your code and just push it to production, and we take care of all the hard stuff.

>> So you're broke-

>> Scaling, load balancers-

>> Yeah, all the clustering-...

>> , all that stuff....

>> and all of that.

>> Correct.

>> Yeah, and I think that to me is really interesting, because I think that both of us spent time over at AWS and have built stuff there. And again, always building that layer cake. And you always had to build up to get to your platform that you were building. And then you got your platform, but then you had to build the apps on top of that. So, there was a lot of engineering that went into that. Where does this kind of join in with what else is going on at Salesforce as well?

>> So Heroku was acquired a long time ago, 14 years ago, well before my time here at Salesforce. And the original idea behind the acquisition is the reason it's important to Salesforce today. Which is, although Salesforce broadly is very dedicated to no code, low code, easy configuration, all of those things, there's always going to be that time when you need to go and write a serious amount of code. Or maybe you have an existing system that's already built and you need to integrate it into the rest of Salesforce. And so having a PaaS, a platform like Heroku that is very oriented towards what we call pro-code, the professional developer people who write code every day, to kind of extend and add all the superpowers to Salesforce is really important. So, we're an important part of the portfolio for particularly the more sophisticated customers who need to write a lot of code in and around Salesforce. But we also have a lot of customers who use Heroku as their general purpose platform.

>> Yeah. No, I think to me, when I looked at it and I looked at the whole capabilities from the various different languages, it's not like, "Hey, it's one language. It's just Java or JavaScript or .net." It's like Ruby, it's like Go, it's like everything, it seems like that.

>> Yes. Right. We want to meet the developers where they are. And I think you know, you've been in this space, developers love the language and the frameworks that they love. And so, we try to support all of those.

>> And having a previous startup of mine, we actually built on force.com and actually used the platform, so I know how flexible it can be. And help us understand how it ties into kind of that Kubernetes environment, because it has that all in there already wrapped in.

>> It does. So, Heroku was the original cloud native. It was cloud native before cloud native even became a thing. And you might remember when originally the cloud native computing foundation was going to be called the Kubernetes Foundation. And we all decided it really ought to be something more general purpose. And so, there's been a lot of standardization around Kubernetes as kind of the underlying distributed operating system. But Heroku had built a proprietary, a really awesome proprietary version of that. And so, now we're at this point where we are moving to run Kubernetes under the covers of everything Heroku does. And so, we're providing standardization on top of that. And that's giving us the ability to move a lot faster and adapt to a lot of standards that customers really want, things like OTel. This is all kind of deep in the DevOps platform engineering-

>> Absolutely....

>> world. But it allows us for the customers, especially the customers that are using Heroku as their overall platform, those are the things that they really care about.

>> So, it'll help them tie into their observability and their platform engineering and things of that so they can standardize their processes across yourself. Because I have yet to meet one that uses one Kubernetes platform. Most have two or three. And I think they want to run it all in the same manner as well.

>> There's kind of a new wave of folks coming to Kubernetes and cloud native. The last 10 years, Kubernetes just had its 10th year anniversary, a lot of the folks there have been sort of the early adopters, the more sophisticated users. By the way, Salesforce is one of those. We have our internal platform called Hyperforce, which is a big Kubernetes, probably one of the world's largest Kubernetes deployments. But Salesforce has large numbers of people that work on that. And if you're not in this kind of tech-heavy segment or you're not interested in building a huge platform engineering team, hiring all these DevOps, platform engineers, Kubernetes experts, then you can come to Heroku and we'll help you with all of that.

>> So, it's the easy button to help people get up and running a lot quicker on that standardized platform. And I think, how does it really... Because a lot of talk and Mark is just offstage across the street, there's a lot of talk about agent force and AI. And how do you see that tying in with Heroku and what is all being announced this week?

>> Well, so it's sort of like what we were talking about before. If you're building agents, but then you have some of your own special sauce, proprietary code, things that you want to do to extend those agents, give them some superpowers, do something really interesting that requires you to write a lot of code, you want an easy way to do that. And Heroku was that way. And we also view that all applications are becoming AI applications. Now, that doesn't mean that every application is itself a model or an LLM. But every application has some need to connect to inference model, or to provide those kinds of services, whether it's RAG or access to some kind of easy inference model.

>> Yeah, it would seem, and I think again, you have some database services that can be deployed, you have all of the different pieces. So, people can come in and almost make their, what would be either a Helm chart or whatever, their configuration chart of their application and bring all those... And you must make that easy for people to go and do.

>> We do. One of the bits of Heroku magic is that we bring not just the compute pieces like Kubernetes does, but we also bring into all the database services, and Kafka and messaging. So if you're going to build an application, it isn't just building some code and deploying some code. You also need a database and you need a schema. And you need to securely connect the applications to the data. And you need the load balancing and all of those things. About half the Heroku business is running databases. It's a really big business. That's why we introduced pgvector support. If you're using RAG or some of these other techniques, you need to have a Vector database. And so we've been providing that now actually, for quite a while. So, that's another way that we've been supporting these kinds of AI superpower agents.

>> And can they bring their own database as well or their own models? It would seem like that.

>> You can. You can. And it's pretty common for folks, especially in Heroku, typically tend to be AWS customers as well. But maybe they do something with a GCP database service. So you can connect those things, and we make it easy to connect to additional services. How many database services does AWS have? Many, many, many.

>> It depends if you count Snowflake and Databricks and all of those on top of it.

>> On all of those. So, there's always a need to connect to those things externally. But we're also always looking for, where is the... We call it painting it purple, and that's our way to say, "Oh, it's part of the kit.", as opposed to being outside the kit. And so we're always looking for ways to extend that as well.

>> And I would assume being that it's a platform as a service, that they're using the databases that are painted purple already, it's super easy and all of that.

>> Super easy.

>> The backup and all of the different high availability is already built in for you.

>> It's all built in, it's super easy. The documentation is great, the support is great. So all of the pieces that you need together in the platform are what we do, and we do it really well.

>> And I would assume, like you said, you started talking about connecting up to other services like at AWS or something like that. Obviously, Salesforce has MuleSoft as well. And that must play some role in that.

>> We do. In fact, we have some really great customer talks going on this week. Lamborghini is here talking about using Data Cloud and Heroku together. Goosehead insurance is here talking about using MuleSoft and Heroku together.

>> Oh, great.

>> And so yes, if you want to use the... There's a lot of customer overlap between Heroku and MuleSoft, but not a lot of technical overlap. What MuleSoft does really well is not the same things that Heroku does really well. So if you need those two things together, that's an awesome match.

>> Yeah, just knowing Mule Soft, like if I want to connect up to a mainframe to bring data into an application that is on Heroku or being built out on Heroku-

>> Perfect. Perfect ....

>> it would seem like a very good way to bring that together 'cause it gives you that gateway. What else are you seeing? Again, you've been around Kubernetes for a while. How are you seeing the environment? Just the whole CNCF like you talked about, and not calling it Kubernetes Foundation, which I agree with. Because I think it's more about cloud native now, because like you said, 10 years on and with Kubernetes, it's solved by people like Heroku and others that really do it well.

>> I think the thing that I still see customers struggling with, and it's becoming even more important in this new AI era, is the ability to do, if you say CI/CD, can you continuously test the thing? And then can you continuously deploy that into production? That's really hard. Most people, I think, do a pretty good job on CI. That's pretty well understood. But the ability to build all of the pipelines that you need in order to make a change and get it into production without a human touching, it is a really hard chasm to cross. And a lot of companies haven't yet crossed it. That is one of the ways that having... So Heroku isn't just data database and compute services, it also includes the tooling that you need to run all of those pipelines. And the pipeline support is super critical. It's becoming even more critical now because if you think about agents building code, the effect on the team is very much like you just added a whole bunch of interns to your team, all writing code. All of this code needs to get to production. And if you haven't built the organizational muscle, you don't have the tooling in place to absorb all of those code changes that need to get to production, you're just going to be swamped in the soup of, "Oh, well, we have all this cool stuff, but how can we be sure that when we deploy it to production, it's still going to work?"

>> So, how does Heroku help with that CI/CD pipeline? I think you brought up testing for instance, and we were talking to one of the ecosystem partners about that just yesterday. And like you said, one of the critical things about AI is, how do you limit hallucinations? And part of that is the model training and other stuff that goes on in fine-tuning. But part of it is also making sure that the prompt engineering and all of that, which is all code at the end of the day. How is Heroku helping developers in general with that CI/CD?

>> So really the main point of CD, or I would say even CI, is to enable teams to go fast and not break things in production. But part of the go fast thing is, "Oh, I made a change and it would break something, but the tooling keeps it from getting into production." And so, that enables the teams to really go fast. And so, what's the effect of hallucination? The effect of hallucination should be, "Oh, this test case didn't pass.", which this is hard. This is very hard stuff to do, and this is the new frontier. But the final line of safety between the hallucination or any other mistake that a human makes or an agent makes, is having that great layer of testing in between. And so, you need a lot of tooling. If you're going to build a platform on top of Kubernetes, you actually have to go and figure out all of the pipeline tooling as well. And select from a vast number of different tools in that CNCF landscape, which there's over 200 projects now.

>> Yeah, it's unbelievable the projects and the mountain. There seems like there's a new one every day.

>> Exactly.

>> And which ones are in sandbox, which ones are actually being promoted, and things of that.

>> Each one of those things is sufficiently complicated that you could have a team of people that each one works on those things. And you still haven't integrated them all.

>> Right.

>> And so if you really want... So it's sort of the easy button in one way, but it's also the safe button in another. It's really important that when you pull all those things together, that you keep them operationally stable. One of those things changes, did it break everything else? There's a security vulnerability here. Did you even know that there was security vulnerability that needed to be patched? Heroku takes care of all of those things.

>> Right. And I think by saying what you said around Salesforce, one of the largest users of Kubernetes across with... The underlying, we actually were talking to David a little bit about that yesterday, around how that's actually built out on other clouds and things of that nature as a platform. It would seem like this ties in really nicely with the experience that Salesforce has running at scale as well.

>> Salesforce, if Salesforce were an external customer of Heroku, it would be our largest customer. And so, Salesforce uses Heroku in so many things. Trailhead, for example, runs entirely on Heroku, the BT org. That would be called the IT org, the CIO org in other companies, uses Heroku to deploy and run thousands and thousands of apps. I think it's over 10,000 apps now. So, Salesforce uses Heroku for this capability at scale.

>> As they say, I know it's not popular, but dog fooding it and understanding, and eating your own dog food kind of thing.

>> It's so important.

>> I think the better way to say this is sipping your own champagne.

>> Yes.

>> It's a better way of saying it. But that's always interesting. Well, hey Bob, I really appreciate you coming on. I think this really helps people understand just the expertise that Heroku and Salesforce bring to bear. Especially when people think of this, they can think of Kubernetes and they can think of Salesforce in the same terms.

>> That's great. Thanks so much. I'm happy to talk DevOps with you anytime.

>> Oh, we will. I'll take you up on that-

>> All right....

>> that's for sure. And thank you for watching this episode from Dreamforce 2024. We're live all day here at the NYSC. Stay tuned for more. And we'll be right back.