PurePerformance - Dynatrace Perform 2017 Tuesday Wrap-up

Episode Date: February 8, 2017

Wrapping up the day with interviews from Henrik Rexed from Neotys and our new friends Rajesh Jain and Maggie Ambrose from Pivotal Labs.

It's time for PerfBites. What the f*** is PerfBites? The fourth square meal of the day. Don't bogart the PerfBites. F*** waffles. Microwave ready. Add nutritional value to your brain. I'm Jackaloo.

It's time for PerfBites with your hosts Mark Tomlinson, James Pulley, and Howard Chorney. PerfBites. Whatever. How is everyone doing? Good. This is Brian Wilson with... What's my podcast? Pure Performance. And we have Mr. Mark Tomlinson

who's a little busy, but he might be joining in. Who's playing the straight man today. I was busy being a performance engineer. I don't know if you guys know what that's like. I forget what that's like. Sometimes we have to do it. Heinrich Rexit of Neotis. Hi, everyone.

So, Heinrich, welcome to PerfBytes. Yeah. And not only PerfBytes, you're also on Pure Performance. Right, we're simulcasting. You're on two podcasts at the same time. Wow, I'm impressed. I don't know if I can handle that stress. And that is, in fact, legal in the state of Nevada.

Yeah. Yes. Yeah. Some states, it is not legal. Utah, maybe, as well. Yeah. Yeah. Of course, Utah. So, hello Yeah. Of course Utah.

So hello. Nice to meet you. That's a really great pleasure to be part of Perfbyte and Pure Performance. For us it's the first time for Neodys. And we want to make some extra shows

in the next coming month. So to let people understand a bit more about who we are and what we do on the market. And for the PerfBytes community, Neotis has just joined as a sponsor of PerfBytes News of the Damned. So you should expect a full feature story or podcast on your product line, as well as sponsorship showing up as part of the News of the Damned podcast. Do you like the News of the Damned? Yeah, I mean, it's a great honor to be part of this. That would be a way to at least have feedback from our product

and share news about our product as well. Yeah. Now, you realize Satan introduces News of the Damned, right? You knew this? You know about Satan? Yeah, a little bit. A little bit, yes. He's not joining us at this moment.

No, he's not. In terms of Neotis' background and an alignment with News of the Damned, we're really showcasing crashing terrible websites, of which you guys have a tremendous amount of knowledge, like any of the people working day in, day out, helping customers. In your recent memory, what is the craziest web performance problem, system crash that you worked on in Neotis' world?

I would say it's more about mythology problems more than crash. I mean, it was a crash, of course. But it was during... So you're thinking bad test and then you come in and clean up the process related to that? Yeah, or probably people didn't realize that you just missed a piece of the load

just testing one small area of the application without taking into account the browsers and bigger parts of the load. Just testing one small area of the application without taking into account the browsers. Bigger parts of the simulation. Exactly. And that was a big problem. It was during the European Champion League. Yeah. It was for a betting company.

So you can imagine that they were expecting a lot of user betting during this event. Yeah. And so they didn't test really well. And they just realized after the first game that, oh, my God, there was more mobile user than expected. They thought that people would just bet from the browser. Yeah. So what we see on the mobile side of the house is an increase in what you would call latch time. That is, the network is dirtier, so it takes longer to attach,

it takes longer to hold the conversation, and longer to detach, so there's a longer hold overall on stack resources, on the IP stack. So you can crash a web server with a smaller number of mobile clients than you can with an equivalent desktop number, which is kind of interesting from a performance perspective. Yeah, or typically they probably underestimate the number of nodes for the mobile users. I mean, the rest of the business logic is common piece, but the node for the mobile user will underestimate it.

So they didn't imagine that there would be so much spikes with those users. So typically, the user wasn't able to take bets during half games. So there was a lot of money lost for those companies. I did a little bit of database work with the online crazy betting back in the day. Some guys out of Vienna, thinking of another one, I think it's in Estonia. You think in Estonia, they do really cool things, right? There's some crazy things potentially in Estonia as well. One of them was crazy online betting, you know, masked behind other stuff. But I digress.

There's a crazy story about backup batteries in the middle of the night trying to keep the betting site up and running. Yeah, forget it. You think they can afford a generator if they're a betting site? They couldn't find a generator, but they're like, we have just enough time to take all the batteries out of this site and put them in the truck and drive over to the other DR site and swap the batteries and keep it going until the lights came on.

That's just nuts. It is. But they had the power to do it. Get it? The power. Battery power? Oh. Awesome.

Yes, I have to live with this guy. I know. So, yeah, welcome, Neonis. Tell us what's new. What's new? Neonis World. So, recently, we are more focused at the moment on automation. Because DevOps is a big trend at the moment, for us it's very important to bring

a lot of features to help our customers to be able to fully automate and include load testing in the DevOps pipeline. So bringing a lot of APIs, bringing a lot of small features to automate, change the monitoring stack of the project, be able to change which load generator you want. So if you have a continuous delivery or continuous deployment process where you include load testing, those features are very important and help you to fully automate load testing. So I'm not thinking about system-wide testing here. It's more about testing in an early stage. Of course, when you have to do system-wide testing,

you have a lot of also features around it. But at the moment, we really focus on this area, on automation. So what are some of the features that are coming out? So recently, we've come up with the selenium conversions, which means after the build, you can run some G-unit tests So recently we've come up with the Selenium conversions, which means after the build you can run some G-unit tests on Selenium, and if it goes well, you can convert those testing assets into NEO scripts directly. Right. And then because last year we came up with user path maintenance,

which is a way to reduce the maintenance task of load testers. The ideas of that feature is like comparing the recording and the existing scripts and merge it. So if you have extractors, if you don't have logic... Sort of automating the diff from release to release to release. Exactly. So you keep versioning along with all of the diffs, so you're able to go back to version X?

We keep the history, but the idea is if you have, for example, extra HTTP calls that has been added or if there are a couple of variables that have been added. Missing field here and there, yeah. For example, then we just merge it. So if you already have correlated and added a lot of variables within your existing use case, all those design aspects that you've done in the past will be automatically placed in the new recording.

So this is huge because normally the labor factor for rebuild of an existing business process that you know is one half of the amount of time it takes to do the original business process because you know all of the dynamic elements to handle. You have all the data. But if you're reducing that even further by just simply integrating the extra hidden fields, the extra resources attached, things of that nature, that must reduce your cycle time to, what, a quarter or a tenth of the amount of labor to rebuild? A couple of customers are saying that they're saving like 50% or 60% more, depending on the customer, on the maintenance. So that would be dropping from a 50% labor overall to 25% or even less.

Yeah. So a couple of things. First of all, that sounds pretty cool. And I think you were trying to ask, are the testing assets also in version control, in Git? So at the moment... Can you do that as well? Yeah. So a couple of years back, we wanted to bring a lot of connectivity with SVN, which was good at that time.

But then Git kind of took off. Yeah, it took off. So at the moment, we are be aligning with Git. But we want to bring those features where we can detect the commits made by different testers. What we had in mind initially is you have the ability to have several testers working on the same project at the same time and be able to commit and merge those assets.

Yeah. So this is what we've done in the past and we want to add the Git aspect because Git is so much popular. If you're not in Git space today, I think you might be left behind. I'm not cool. I'm not in Git.

You better Git going. Git it done. So I do have another question just on the diffs because if you think historically, I'm thinking of doing old diff scripts in Lode Runner and we had, I think, WinDiff we used to ship in the product. I think it's still in the product. It's maybe still in there. Probably an old version left over from a long time ago.

Yes. But that was, you know, for everyone, in different tools included, that was sort of the manual step to diff things to look for correlations. So I record it once. I can record it a second time and see which fields were different as I'm working through the correlation nightmare. I'm wondering if the same sort of automated, diffing kind of merging is something that can help with correlation as well.

I think if you look at it, I mean, most of the people, when they design, the first time you design, you put a lot of effort on that because you want to add to the logic. You put the assertions. You put the if and else. You control the errors.

So what we think is important once you make that effort, then if you have a new version next week after, the changes of that will be very small. So maybe a couple of parameters added on the form or maybe an extra request needs to be added. So let's take advantage of what has been done by the first version of the script and

to reduce the time to maintain. To find the changes and maintain. I remember 10 years ago it was sometimes less expensive to re-record and re-do the design rather than do the maintenance of the script. Absolutely. Especially if there was a personnel change.

You know, because like any code, you hand it back to the original developer. And let's face it. If you're not following good development practices, no comments, it's not error handling, you don't have good variable names, someone could pick up your script your test code and say what the heck is this? And you're absolutely right Heinrich in that case just rebuild it from scratch

because I can't take the time to figure out what this person is doing. Well this gets to your point of being able to actually gen the script conversion from Selenium. So you could have some people maintaining Selenium and even if there were no changes to the Selenium script, you could still generate a new lower-level script from that.

Yes. So that actually, the re-recording is kind of automated in the conversion process as well. Yeah, that was the idea, is to fully automate a process where if it has minor changes, you can imagine where you build, you deploy, you run the GUnit in CELIM format, and then you convert it.

And if it's fine and you have a check from NeoLoad that everything goes well, then, yeah, okay, let's go to the next step, which is I'm going to run the load test. Yeah. So we're at Dynatrace Perform. The other thing I should ask you, I think appropriate to this, is what's going on integrations-wise for being able to do new stuff with Neotis, Neoload, and Dynatrace?

Let's be blunt. You're at Dynatrace Perform. Why are you here? Yeah, we're here because we're a gold sponsor for Dynatrace. Dynatrace is a great partner with a great product. Oh, shucks. Isn't it?

Brian is blushing at this point. So we had this integration since a couple of years back. I think because people are not aware of it, it's still interesting that it reminds what we have today. The ability to, of course, like most of the load testing tools are already doing it.

We send it text to Dynatrace, so it appears like a synthetic traffic. But we also take advantage of the API delivered by Dynatrace to start a session, stop a session, so you can have recorded sessions available directly in AppMob. Yeah, so automatically giving you a stored session starting at the beginning of the test. That's very nice. Let me ask. Sometimes if you automate a load test and it's sort of unattended,

you just always store the session because you can come back and look at it. Right, yeah. Yeah, yeah, yeah, totally. Go ahead. Thank you, Mark. I was just curious because one of the features I never got to use when I was a Dynatrace customer because I don't know if it had ever been brought to my attention was the session comparison. And I was just curious if you see people using that concept of Dynatrace after they're running their tests.

Or are you even familiar with the whole session comparison piece? Yeah, I'm familiar. It makes sense. Again, it's exactly like comparing two test results in a load test. Sort of a baselining feature? It's different than baselining. Yeah, different than baselining.

Basically, what it's going to do is you can set your previous test versus your current test along things like method execution times, database executions, and all that's going to show you what improved either in counts or response times, what degraded, and then different variations of gray might show you what's been introduced or deprecated from the build. So let's say you get a new build.

So now we get a new build. We pop it in. We run it. And now we've got an A-B comparison. Right. And if you just see on that first page all red, fail it now and then dig in to what's going on. Anyway, go ahead. Yeah, but that would mean that you need to have the same context, the same testing context.

I mean, if the application is, the environment is totally different, you cannot compare. Or the workload or feature changes significantly enough. It may, yeah, they shouldn't be comparative because enough change has happened. But yeah, for like repeated continuous testing with very small changes, it still probably gives you a good idea. I would look at something, say, the layer

breakdowns and such where you could see these layers should never change because I didn't change any of that code, but we added something for a cryptography call or a little overhead in some other transformation or something, right? So at least you could compare and then go into a layer view on the comparison. That'd be

interesting. But I do like the start-stop session. Yeah, that's always a great thing, those API for, you can even name it in the dynamic, you can register the test run in the test dashboard and all that. There's a lot of really cool things you can do with that API. Yeah, because there's nothing worse than coming back after a long weekend where automated tests have been running

and your data is gone because it's fallen off the end of the buffer. Yeah. You just need a bigger database. No, that's disk. Oh, that's disk. Well, database is disk too, but storage. Performance data warehouse. So let's give a shout out to Seagate on their 10-terabyte drives.

There you go. Perfect. There you go. 10 terabytes. It's SSD, 10 terabytes? That'd take quite a while in a load-testing environment to fill that up. All right. Tell us, now, you particularly, as a role, what is your role at Neotis? Is it specific to performance?

Yeah. So I started like a performance engineer. I used to be the technical evangelist. Now I changed this year to build up a joint solution with partners. Okay. So, for example, with Dynatrace, I really want to provide a new version of that integration with new features. Great. And similar thing with

other partners. On IoT, we have a couple of ideas. So the idea is to be able to either customize the application, our product for large GSIs or service companies that could address

specific application that is not covered by default by a product. Right. And also bring value. I mean, if you customize the product and put it in sort of marketplace, people will love it. I mean, bring new features without waiting for the new version of a product. That makes a lot of sense. And like you say, some of the older techniques like tagging in the headers and such,

there's all sorts of new things coming, like from Dynatrace. Like, can we make use of those? So it maybe even isn't very hard from your side of the equation. It's just, wow, now we can make use of all these APIs and just sort of grow the integration ecosystem for the products. We're mostly interested in NeoLoad, I think, mostly from a performance standpoint. Yeah, yeah, that functionality stuff. You know, if it doesn't work when it gets to us, from a performance standpoint. Yeah, yeah, that functionality stuff. If it doesn't work when it gets to us, we don't care.

Yes, that's true. But it's not, actually, because you're sort of leading into having a Selenium script that's doing functional testing. It maybe isn't doing a functional test, but it is a representation of the business flow. You're sort of blending the functional automation into the performance automation and could be done continuously. Is there also going into operations? Are there things you do for synthetic monitoring or anything like that in Neodys? We tried it in the past.

And I think on this market, there is so much competition. Yeah, tons. I think we won't be able to bring as much value than a great company like Dynatrace. So I think it's more interesting for us to collaborate with them rather than starting to try to compete with them. So for us, it's more about taking advantage of what Dynatrace or others are having as features and having more about the shift of the approach. I mean, they are measuring everything on production. They are taking a lot of things that could be very useful for a performance engineer. Let's say we've tested on Black Friday and I want to redo the tests.

So why don't I just extract what happened on last Black Friday and create automatically the test cases. That would be a way to reduce the cost of load testings. Yeah, and give you some flexibility to not have to remember it all right here in my brain.

In my brain. Sorry. We see Mark's problem, don't we? It's crazy. It's very, very crazy. So that's awesome. What else should we ask Heinrich? Well, we've asked everybody who sat in your chair, what are the coolest things you've seen so far at Dynatrace?

At the show. Hey, the bells are going off. The coolest thing is, I would say, the coffee. No, no. That should be the hottest thing. No, it's the hottest thing. No, I would say that

there is a lot of things. I wasn't able to see everything, but I think there is a lot of escape room in the bar. I would like to see that. The Xbox, also 360 contests. There are a lot of great demonstrations of the product.

So if you want to see new features, it's a really good event. Yeah. What did you think of the Davis demo? It was the demo live effect, you know? Yeah, that's right. The lighting, the stage, the live feature. It's exciting. When Davis has a hologram that shows up at the same time, be scared.

I'd be scared. It's like the Wizard of Oz. As soon as it goes into cyborg robotics and the robot Heinrich walks in and solves all your problems. It reminds me of the movie Her. The robot Heinrich walks in and solves all your problems. No, I don't think we should go there. It reminds me of the movie Her. Have you seen it? Oh, yeah. I started watching that. So maybe once somebody will fall in love in Davis.

It could be. Could be. It could be the first. Maybe it'll be you. Who is the voice in Her? It was Scarlett Johansson? I think so, yeah. So how come we don It was Scarlett Johansson? I think so, yeah. So how come we don't have Scarlett Johansson's voice as Davis?

Because we'd have to pay a lot for that. We'd have to pay a lot for that, yeah. I think this is a computer-generated voice, as you can tell. Who is the actor in that? Marky Mark. No, no, no. It wasn't Wahlberg. It wasn't Marky Mark?

No, it was Joaquin Phoenix. Oh, yeah. It was Joaquin Phoenix. I don't know why I thought it was Marky Mark. Yeah. It's a little bit of a crazy thing, but I do like him as an actor. He's rather nuts as an actor. What was in Gladiator?

Remember Gladiator? I didn't care for that movie too much. He was in that. Yeah. And he was like the crazy. He was the mad emperor. Well, yes. The son of the original.

I enjoyed Inherent Vice. That's one of his Thomas Pynchon books. Yeah, there you go. This has nothing to do with performance. No, not at all. Now we're talking, this is like the Joaquin Phoenix show. That's exactly. Yeah, so that's fantastic.

What else would you look forward to as being a sponsor of News of the Damned and PerfBytes? What would you like to see us do more of? I think for us it's the opportunity to announce some really great features that we have in the coming year. Okay. Also to share experience that we have because we have a lab internally where we are trying to test a couple of features. Some of them are in the market. Some of them are in the basement of our headquarter. So I think it's really interesting

to share that with your community. Yeah, so maybe we could do like just PerfBytes listeners could be like We've added a new segment to all of our PerfBytes podcasts. I think you're welcome to hop on anytime with news

from Neotis. And we could get Peer Performance to do this as well. We could be a virtual customer advisory board. That would be great. Don't you think? We could have it on the Slack channel. We have a Slack team for PerfBytes, so people could join the virtual cab. I'm uncomfortable with the concept of me being virtualized.

No, it has nothing to do with Scarlett Johansson and Joaquin Phoenix and the weird AI robot Davis thing. Okay. Although we could do that if you really want to. We can make a model of you. I think the world is a better place if I'm not replicated. I would like to combine it with miniaturization so you could have like a mini James robot I think that would be a really bad idea a lot easier to travel

just toss them in the drawer exactly I do want to give a shout out to back in 2009 I believe it was my former job did were looking have computers back then yeah back yes they did this was back in 2009 we were looking to uh to unseat a certain dominant load testing tool that mark used to work for and uh we ended up not finding any replacement but i

remember evaluating uh there was there was a distributed issue with Neotis at the time as far as being able to distribute amongst different generators, but one thing that super impressed me was I downloaded it, I recorded, and without even touching the help manual, I did a correlation and got it all to run.

It was super, super simple to adopt to as opposed to some of the other tools out there. Batteries are dying. So anyway, it was really easy to figure out, and I've always had respect for the tool since then, even though we didn't go ahead and purchase it. And that tool was OpenSTA, right? The entrenched player you were looking at.

I have never heard of that. QA load? We use WinRunner to generate a lot of load. Oh, okay. Anyhow, that's all. One thing I think would be interesting to figure out. Sorry, one thing for you. Neotis in the cloud.

Yeah. That's something I think I don't even know that much about. Maybe make some comments just about how you guys haveotis in the cloud. That's something I think I don't even know that much about. Maybe make some comments just about how you guys have product presence in the cloud. Because there's so many load testing tools and things that are running from the cloud. In fact, people don't consider us like a cloud testing solution. But if you look at our offer, it's been almost six years now we have a cloud platform. What we pushed on the cloud was our load generators. So what we wanted is instead of forcing our customers to have an Amazon AWS account or a Rackspace account or whatever it is,

we said, okay, you want a general load on the cloud, we have a cloud platform. We are not cloud providers, but what we are delivering on the market is an engine that will spin up automatically machines on several locations. So that is a really great advantage, a really great feature, because we did last year 2 million concurrent tests for

an entertainment, it was a TV show in fact. Oh, that's cool. And we needed more than 500 load generators. Yeah, I have that in my basement. I have that many virtual machines on one box. Oh, Jesus. I don't want to mention what's in there.

They run really slow, but they're all there. That's right. You know our, I guess it there. They run really slow, but they're all there. That's right. That's pretty impressive. You know our, I guess it's a competitor, sort of, but Tim Koopmans from Flood.io. Yeah. So Tim also, I think he gave a talk about doing like a million views or something like that. So we actually should keep a leaderboard of who's run the largest load test.

Normally, I think we have a partner this year that is planning a 3.5 million concurrent tests. So let us know how this goes. Yeah, and we... Let's have the little bar. Yeah, we want... We would like to have the agreement with the customer. If he agrees, we probably want to make some small videos during the test. Sure.

And push that on the web. That would be very cool. If you can get a capture of an actual machine that goes up in flames, that would be great. Yes. That would be the most fun. With the smoke around it. Smoke coming around, the melting plastic, that smell.

You know that smell, that melting plastic smell? That acrid green goo type smell. Yeah, the metal soldering welding smell. Yeah. That's what we're looking for. It's got kind of a copper aftertaste to it. Yeah, exactly. I think we described that pretty well for being on the radio. Yeah. Yeah, exactly's what we're looking for. It's got kind of a copper aftertaste to it. Yeah, exactly. I think we described that pretty well for being on the radio.

Yeah. Yeah, exactly. All right. Now, Heinrich, I hold out hope at some point that there is a test tool manufacturer that will integrate an Age of Empires overlay for test execution. So the application under test is represented by a bunch of castles, and the virtual users come up and start poking the castle. Can we get some animation happening?

When it gets under stress, it catches on fire and stuff like that. We all load skins. The thing is, at the moment, our developers are focused on features, but it would be very cool to have sort of a game, a hiding game. Gamification is very popular. So think about this. Executives who watch a performance test, have you ever seen their eyes glaze over faster? Yeah, they go comatose.

They're like, I see lines. What am I watching? Sometimes it's when you make the lines red and the red lines are going down. The CFO just freaks out because they're like, flashbacks of the market. What does that mean? Redline's going down. So, yeah, so if we can have little soldier characters and sound effects of, like, when the web server starts to die,

you hear it whimper. I'm sure that they're going to get right on that, right? You're putting that to the top of your RFE list. I know it's a fantasy feature, but I always ask for it. But it could be gamification. You could get young kids that are totally into games. Five-year-olds. Five-year-olds.

You can get five-year-olds. Start getting five-year-olds into performance now. And that is an objective, I think. That would be great. If the tobacco industry can do it, we could do it for performance. I think so. Yes, that's right. I'm going to try to push my kid to see if they're able to do it.

If she does. So, Henrik, I apologize. I put you on the spot. Total sidetrack. Yeah, I will ask the product management to put that on the next release for sure. Can you at least leave it on the backlog, even if you never do it?

Just so we can say it's on the backlog. Yeah, that's actually a good idea. I'll probably invite the product management on the next show. Absolutely. Age of Empires animations in the product. Yeah, the animation skin for test execution. Yes.

Awesome. Heinrich, thank you so much again for your sponsorship and mostly for your participation here today and supporting Dynatrace. It's great to meet you in person, and this is going to be really good. Thank you so much. Pleasure to meet you.

Thank you. Thank you so much. Pleasure to meet you. Thank you. Thank you, sir. Pure Performance listeners. Yes, PerfBytes and Pure Performance together. So it is, even though we are broadcasting live with you, the episode stays out there forever. So you can go listen to it. You can send it to your family, your friends, and say listen. Or your grandchildren someday.

You can say, Grandma, what's a podcast? Oh, well. Back in my day. It was the new radio. What's radio? I don't know. We like to be

hip and new and contemporary with this. What do you call this thing? It's a podcast. The internets. I think it's a podcast. We're here with Pivotal Labs. Right. Awesome.

Maggie. Rajesh and Maggie. Hello. Hello. How are you guys? Good. Great. Thanks for having us.

How's Dynatrace Perform treating you so far? Dynatrace Perform has been a wonderful event. Really? We've gotten to meet a lot of great people. Lots of great Dynatrace Perform has been a wonderful event. Really? We've gotten to meet a lot of great people. Lots of great Dynatrace people. Lots of great customers. Cool. Food's good? Food is delicious. Okay.

Snacks were great. I would add, actually, I mean, I think the keynote was very impressive today morning. Yes. What we heard and what Davis, Alexa or Davis, they can do. That was really impressive. Good show. And I got to say, just so the listeners know,

Rajesh used to work with us at Dynatrace. So it's funny because I've been aware of all the stuff going on, obviously, here. And in my mind, you kind of are aware of that stuff. That's all been going on. But obviously, you're not because you haven't been here in a while. So hearing you like, oh, wow, it was really cool. It is cool stuff, but that just kind of struck me as funny because I thought in the back of my head,

well, of course you knew about that, but you didn't. But you didn't think it was like Skynet for Terminator, no? Really? Yeah. A little intelligent guy inside the computer? It was impressive. I thought it was kind of a made-up video. I didn't know it was live until the last part when they did actually a live demo.

Then I thought, oh, wow, this is something. You thought we were going to fake it, huh? It's not fake. I thought it was faking. For the first part when John played the video, I thought it was faking. But then when Alois did the actual demo, it was live. That would be really bad if we faked it. That would be amazing.

There you go. I think that would be really bad if we faked it. That would be amazing. There you go. I think that would be the end of us. Pay no attention to the man behind the curtain. That's exactly right. They could have pulled a Milli Vanilli and the CD would have skipped. Right? Yes. Lip syncing. So Pivotal.

Great to have you. Thanks. What is new at Pivotal? Yes. What is new? There was a... There was some announcement a little while ago. You were going to let them talk. I was. Talk. Yeah. So... It's all the coffee. Pivotal, actually, I mean, if you ask us what's new,

it's like what's new this week is the right question because what's new this week is going to be like old next week. Okay. We move at such a high velocity. So then for each of you you what's new and exciting to you what what did what are the new things in the last few months last few weeks just last week yeah last feels and it's tuesday so it's pretty much something that i think is is kind of interesting um something i heard about recently last week is just more investment into or more focus on the automation story with Pivotal

and with automation of getting things, getting your Pivotal tiles and Pivotal Cloud Foundry installation upgraded and doing that kind of automagically in such a way. That's actually, yeah, to add on that, that's very impressive. That's something where we at Pivotal as a company are investing a lot of money because what we are seeing is most of our accounts are not smaller companies, like huge enterprise companies. Right.

And when we have now it's like year three of Pivotal, and we are discovering that, hey, it's easy to roll out a platform, right, the first thing, but how do you actually keep it updated and how do you actually maintain it and how do you make sure it's always current do you actually maintain it and how do you make sure it's always current. So we are investing a lot of money in that and we have what we call our customer zero which is basically the automated

way what Maggie was talking about. We are making a lot of investments and hey, let's actually make sure that a lot of automation built into it from the right get go so that as you not only are pushing your apps, but you're pushing your platform, it's all automated. You have some fans walking by.

That's pretty awesome. I've never gotten a shout-out for being in any company that I've been in. There you go. You've got to go work for Pivotal now. And with that, making it all very repeatable so that it can be repeated across all the different platform installations. And then also when you're talking about what's new, it's like we release software once a week or twice a week sometimes. And for some of our customers, that's very...

It's crazy. It's crazy. It's overwhelming. They're just going to have to get used to it, though. So it's the new way. Well, it's the new way. We'll try to help. What about... And I'm not sure. You're just going to have to get used to it, though. It's the new way. Well, we'll try to help.

What about, and I'm not sure, maybe just give us a little high level, like people don't even know who you guys are with Pivotal. What do you do? Just kind of high level. Yeah, so Pivotal has been in existence for the last, I would say, four years now. We call us startups, normal startups. What we do is we actually have a platform that we provide on top of your existing IaaS.

So you might be running your workloads today on AWS or in Azure or in GCP. But what we do is we put a platform on top of it so it becomes easier for someone to take those workloads that are running on GCP and move to Azure. If you get better money, value for money from that particular vendor,

let's move that over there. So we actually kind of abstract that underlying infrastructure and make it very easy for you to move your workloads based on your best possible location. It could be based on cost. It could be based on security regulations. It could be anything else that you think about. And with that, so that's one part of Pivotal,

and that software is called Pivotal Cloud Foundry. It's an open-source organization. Pivotal is the biggest contributor to it, but we also have other vendors who are contributing to it, actively contributing to it. It's growing. It's growing rapidly. We also do a couple of other things that we do. One is our labs, Pivotal Labs.

So Pivotal Labs is a place where we teach people how to write better software every day. Again, there are some philosophical things that we think about, like how you should be writing software with paired programming, how you should be writing software with high velocity and continuous delivery. Think always about delivering software, delivering value to it. So those are the things that we talk about and teach customers when they come into labs,

get the lab experience, and then they go out and build their own labs. That's what we want to do. We are an enabler. We are not in staff. Yeah. And the third leg of the stool would be also the data suite of things.

So we play in the data space with caching and big data as well. Caching and actually doing, like ETL, moving stuff between cloud data providers as well? Like similar you would do with workloads in the processing space? Yeah, with the Gemfire is the caching solution for that. Right. Very cool. Where does Spring Source, you want to talk about that? Spring Source? Spring Framework that we own, actually, yeah.

Okay. And then another one of our open source projects is the Spring Framework. This is a really popular development framework with Java developers. My guys just updated to the latest version because they were way behind. Technical debt, you Technical debt builds up. Nice. And so just to take a big step back, because a lot of people, as we know,

in the DevOps journey haven't even started. So we have PAS, IaaS, all these different AAS things. Pivotal falls into the platform as a service. Correct, or are you going higher? Not squarely, yeah. Yeah, so we're changing the argument. We are saying that the PaaS world is very narrow, platform as a service.

We talk about as platform, right? So we actually, platform encompasses about, hey, your runtime to run your application, which is your PaaS, but also how do you abstract out the ISs and how do you actually also run different workloads. Maybe it could be data services or it could be your spring apps. There's a whole slew of things that you can run in the platform.

So it's just not the PaaS, narrow PaaS, but it's the whole platform. Right, right, okay. Maggie, you were mentioning automation before in terms of kind of cool stuff happening. Where does that innovation kind of fit in with the bigger picture that we just talked about? In terms of the Pivotal bigger picture? Yeah, yeah, yeah. But like automation,

I'm thinking CI, CD. Yeah, yeah. So, on top of all of the things that we also just mentioned, Pivotal's also always doing different open source projects. So the automation is around an open source project called Concourse. Okay.

That's kind of our project in the CI, CD pipeline. Great. It's an open source project. A lot of our customers are using that for automation of the platform upgrades themselves. And Rajesh loves Concourse, so he would probably like to add a little bit more as well. Yeah, so I've been talking about Concourse, only Concourse, this whole perform conference. I've not been talking about the little Cloud Foundry.

We've also been talking about Dynatrace. Yeah, and me talking, because I love Concourse, to Maggie's point. Yeah. And, of course, obviously we have the Dynatrace tiles that are available, and they're not the plugger stuff stuff but that's part of the whole idea then with all this bit is collect your proper metrics and do your proper checks and balances before pushing right and overall um with Cloud Foundry the ecosystem is a really big part of that and being able to

integrate to use third-party solutions whether it's around application performance monitoring with Dynatrace we have a great integration but also with um like single sign-party solutions, whether it's around application performance monitoring with Dynatrace, we have a great integration, but also with single sign-on solutions and things like that. Anything the platform can't necessarily do itself, there's an ecosystem to give those things. And for anybody

listening, if they are interested in checking out something, is there a way they can get to see some of the Pivotal in action? Pivotal Web? Is it Pivotal Web, I believe? Pdubs, if you're cool. Pdubs, if you're in the know. Pivotal Web Services.

Yeah, you can go to run.pivotal.io so you can do a free trial and start pushing apps to PCF. Brian did it. I did. I took a little class. That's cool. He was my trainer. Very easy to get something pushed out there. And then, of course, if you look up some of the open source projects, you can jump in

directly from there and bring them in and use them yourself. YouTube channel? Twitter? Any of that? At Pivotal. That's a hard one. That's very hard. Can you spell that for us? Because Pivotal Labs has a separate Twitter, right? That's a hard one. That's very hard. Can you spell that for us?

Because Pivotal Labs has a separate Twitter, right? The one thing to keep in mind is the.io. Sometimes I still accidentally write.com. The wrong place. You'll correct yourself very quickly. But is it a wrong place or is it just the wrong place? No, it's just the wrong place. Okay, I just didn't know if it was one of those situations. I don't know what that is.

No, that's very different. So, the concourse website is a little different. It's concourse.ci. Okay. It's not concourse.io. So, if you want to learn more about concourse, which is, again, an open source tool that, as Maggie was saying earlier, we developed. And that's for what again? That's kind of like a CICD.

We're not paying attention. I'm trying to understand. Is this kind of sitting somewhere in like the Jenkins space? Or what would you say would be an equivalent-esque? Probably, of course, it's much better than anything else out there, I'm sure. But is it in that space, the Jenkins space? Yes. Yes. Jenkins, Bamboo, TeamCity.

Cloudbees does some weird stuff. Cloudbees and Jenkins. Brian, the analogy for you guys, Dynatrace will be like, is it in the same space as Wiley? Exactly. That's what I'm saying. IT cam.

How can you even bring that in? Not exactly. When you said Bamboo and Jenkins. All I said was Jenkins. Come on, Jenkins got the market. IT cam. How can you even bring that in? Not exactly. You said Bamboo and Jenkins. All I said was Jenkins. Come on, Jenkins got the market. I know. Is there any kind of free trial with that? It's open source. So you guys can go and...

You can bring it down locally and do it if you want to do it that way. Or run it in the cloud. Or run it in the cloud? Yeah. That's pretty awesome. P-dubs. Yeah. Complete P-dubs. Have. Complete P-dubs.

Have you guys gone to see anything during the conference at all? Did you get to any actual sessions? We got to, I don't know if it was the keynote or the roadmap session. This morning? Yeah, the morning session. Oh, they did the Davis. The really cool Davis demo. Alexa's going to rule everything, I'm afraid, but I'm okay with that.

Alexa, fix my website. It did, but I'm okay with that. Alexa, fix my website. It did, actually. I know. I said fix my website. I didn't know that we could do that. That's pretty cool. That was actually something that Rajesh and I were talking about,

that remediation aspect of rolling back to something that you know is good. It's something that could be built into something like a concourse pipeline to, you know, if a test fails or if an upgrade, if something's going on that's wrong, using a remediation to... And if you remember from our old plugins, we always had those, like the generic execution plugin and the SSH plugin, where you can issue

a command, but it was always a lot more of a dumb level, right? And with the whole Davis thing with this, this is based on a very intelligent factor of conditions when it's going to fire off maybe this rollback or something. Right. It's taking that whole, it's a whole different level. I've got to really find out more about that piece

because that to me was the most impressive piece because I already knew about the Davis stuff and the rest of that. Mm-hmm. Anyhow. Yeah. All right, is there anything else? Are you looking forward to any sessions for the rest of that. Anyhow. Is there anything else? Are you looking forward to any sessions for the rest of the week? Or are you all here for the

rest of the week? We're here. We have Josh McKenty coming. He's giving a keynote tomorrow. He's pivotal. I'm excited for that. That'll be good. We are looking forward.

That's going to be a very I'm excited for that. That'll be good. In Joshua's session, we are looking forward. Of course. He's definitely. All right. Probably. That's going to be a very interesting session. And then I'll be leaving. Maggie's going to be here. We have a few other guys going to be here. Cool. So we're spending most of our time at the booth.

Yeah, in the Performance Cafe. Yeah. Which is good. You can go in and schedule a one-on-one for folks that are here and maybe listen to the show. And you can follow you on social media. But outside of downloading and jumping in and getting your feet wet, there's a forum as well, I believe. I know that Pivotal Labs had a really active forum. Do you guys have a Slack team or any other way to interact and get help?

There is Slack channels that you can get on. So there is a bunch of Slack. I mean, you have Cloud Foundry, open source Slack. Right. Then we have Slack for Pivotal, basically. So Pivotal users can get on the Slack channel. And then we also have Concourse Slack. Concourse has its own Slack organization.

So most of the guys are hanging around in those Slack channels. And I mean, if you have questions, the best thing we monitor internally is Stack Overflow. So if you guys put something in Stack Overflow and tag it as Pivotal, someone from Pivotal will answer that.

We'll jump in. That's pretty awesome. Excellent. Very cool. Well, thank you for your support of Dynatrace on behalf of Brian. Oh, what? Well, thank you for your support of Dynatrace on behalf of Brian. Oh, what? Well, you know, you're a Dynatrace person. Thank you for supporting PerfBytes and joining all the PerfBytes listeners, because that's cool, too, you know? Yeah. That's very awesome.

And, yeah, enjoy the show. Thank you. Thank you. Thank you.