ML-Based Data-Driven Software Development with InfluxDB 2.0

Session date: Dec 07, 2021 08:00am (Pacific Time)

Hari Prasad was using InfluxDB for his personal weekend project as part of his hobby to map the water levels of lakes in his city, rainfall, evaporation index, etc.

His Eureka moment: Software development or any human activity flows with time. He started mapping software development with a time series DB and created an IoT within a software development tool. Before InfluxDB, developing a similar system required huge budget and maintenance efforts. With InfluxDB and its ecosystem, the quality cost & delivery were unbelievable. The Flux addition with 2.0 helped his team with the power of computing. They chose Flux over Python to determine mean, median, mode, and quantiles with Flux’s built-in functions. The talk shares the Templates, Flux queries, Scraper code with the open source community, all of which will be in GitHub where anyone can reference them.

The talk shows how quickly, reliably and cost-effectively you can do data-driven software development by writing custom code. It is so generic that all software development teams from small to large can benefit with little or no maintenance.

Click here for presentation

Watch the Webinar

Watch the webinar “ML-Based Data-Driven Software Development with InfluxDB 2.0” by filling out the form and clicking on the Watch Webinar button on the right. This will open the recording.

[et_pb_toggle _builder_version=”3.17.6” title=”Transcript” title_font_size=”26” border_width_all=”0px” border_width_bottom=”1px” module_class=”transcript-toggle” closed_toggle_background_color=”rgba(255,255,255,0)”]

Here is an unedited transcript of the webinar “ML-Based Data-Driven Software Development with InfluxDB 2.0”. This is provided for those who prefer to read than watch the webinar. Please note that the transcript is raw. We apologize for any transcribing errors.

Speakers:

• Caitlin Croft: Customer Marketing Manager, InfluxData
• Hari Prasad Sudharshan: Leading Automation & DevOps, Fujitsu Network Communications

Caitlin Croft: 00:00:00.400 Hello, everyone, and welcome to today’s webinar. My name is Caitlin Croft:. Very excited to be joined by Hari Prasad, who will be sharing how he’s using InfluxDB. I’ve already seen this once and I’m excited to listen to it all over again. Once again, I know we’re all pretty familiar with Zoom, but down at the bottom of the Zoom window, you will see chat as well as Q&A. Please feel free to post any questions you may have for Hari Prasad there, and we will answer all questions at the end. And the session is, of course, being recorded and will be made available later today. Without further ado, I’m going to hand things off to Hari.

Hari Prasad Sudharshan: 00:00:44.622 Hey, thank you [inaudible] decent introduction. Thank you very much [inaudible]. So we have started doing something called [inaudible]. We call it as Project [inaudible]. And we believe that there are a lot of blind spots in our life and this is an attempt to see, okay [inaudible], the technological advancements in your cars today. So let’s get into this session. Okay, so what is common across all these things that we see around us? Evolution, right? So that’s going to be the theme of today. So I come from India, and the southernmost part of India is one of the metropolitan cities is Chennai. It is home to almost 11 million residents. The city is blessed with sandy, natural urban [inaudible], few hillocks, three seasonal rivers, marshland, and a large lake that stores water for our summer. A perfect amalgamation of natural and modern infrastructure to sustain fast paced life of 21st century, isn’t it? So the city is also home to IT since the 90’s. I’m speaking in front of you all as I’m a beneficiary of these rapid developments. So it’s also home to an ecosystem, which produce noticeable SaaS giants, such as Zoho and Freshworks. So these grew out of this place. The city is also called the cultural capital of archaeological studies, [inaudible] civilization started around - carbon dated numbers are like 2,300 years ago, so it’s kind of both - you still value old tradition, modern as well. So as I told you in the beginning, every city has its own ups and downs, right? So it’s not all rosy. The city has borne the brunt of unplanned explosion and has been affected by rapid urbanization, to sustain 11 million souls.

Hari Prasad Sudharshan: 00:02:59.513 So urban floods have become a norm. Just as we speak, if you Google Chennai now, floods is the first thing that you would get, because just a week back, the city was limping because of floods. And it’s not once, it has happened almost four times in the last decade. Floods that we’ve never seen before. And part of it is natural, part of it is as well the things that we built [inaudible] and what not. These skyscrapers and the multi-level bridges were inundated, the major international airport was shut down for a couple of weeks. And there is a paradigm shift here. Come summer, we have different problems. Would you believe that [inaudible] huge floods runs out of water. Yes, that’s what happened. The city on an average receives an inch more than Seattle. People call it the wettest city, or it’s the rainiest city. And it’s four times more than what LA receives. And to give you a comparison on Europe, it receives twice as much as London. But how can the wettest city run out of water? You see, the bottom right corner, there is a dry lake with a concrete jungle in the backdrop, right? That’s where I used to live and work. And [inaudible] much time. If you see, this is not a simple one [inaudible]. It’s not an outlier, it’s happening everywhere. So if I have to tell you it’s happened in Cape Town, South Africa, Sao Paulo, Brazil, they have similar situations as well.

Hari Prasad Sudharshan: 00:04:48.149 So what’s happening here? And is there a way to solve? You cannot [inaudible], but you can crunch some data and be prepared [inaudible] easy because the open life data is not available and experts in the industry and academia have to pull their thoughts. I think I’m running quite fast here because I just want to get the gist in a very short time. So you have a city which I told you all the rosy things about it and then you have floods. And in the summer, you have the draught. So this is kind of becoming cyclical, where it is becoming a place where it is not easy to live. So can we do something? That’s what we are trying to do here. So I grew up, as I told you, in the city. I have roamed around the city [inaudible], so it’s not like things that we are seeing only now. It’s been happening, and it is going to happen. There’s a huge personal motivation for me in this project. I wouldn’t call it as a hobby project because we decided to address the bottlenecks of unavailability of digitized data. So what we did was - although the digitized data was not available, it was available in paper form. So we have started collecting from various sources since the 1960’s. And we are also looking for a scalable solution. Like most, we started with the natural choice of - we need to store it somewhere, right? The SQL. And we understood one thing, that all these variables change with respect to time. You see, the constants, lake capacity, so these things don’t change, but all these variables like daily water levels. So [inaudible] the water level [inaudible] open the floodgates and when you should close it, helps you in maintaining the sanity of the city, both during floods and during summer.

Hari Prasad Sudharshan: 00:06:49.521 And the rate of change of daily water consumption. We live in tropics, so the evaporation coefficient, the surface temperature, the rainfall received are a lot of variables. All of these variables are with respect to time. So when we started, it was quite difficult for us to map everything on an SQL [inaudible] accidentally, thanks to their search engine optimization. So I was looking for something where I could map with respect to time, and I found a couple of them. And [inaudible] was just easy plug and play. It was during [inaudible]. That version is when we started, and we have now [inaudible] since 1960’s, but still, this is a work in progress. So there are two parts to this discussion. One is this. And we were kind of stopped due to non-availability of data, and that [inaudible] as my weekend project, right? I mean, although I have all my motivation, it does not pay me. So this [inaudible] professional realm. [inaudible] that I’m a DevOps lead. I realized that DevOps are any human activities [inaudible] with respect to time, so I started using Influx for my professional - I could say, hobby projects as well. So what we did is something [inaudible] because we are not making much progress on the floods, but we are now gathering and digitizing data and getting it here. And we have built a few models. And in the last session, I [inaudible] three to four months’ time. We have a plan to open this digitized data to everyone in a form where the academia and the technical community can pull in their thoughts and take it from there.

Hari Prasad Sudharshan: 00:08:50.525 So coming on to the second topic, I wanted to touch first both items. One is okay, how are we using it for a social project, and how did that come to - sorry, my daughter is speaking in the background. How did that change on the professional front? So I’m going to show you the evolution. So evolution is not only for humans - it’s for software as well, right? So there was an idea, “Okay, how would it be if we can map what happened, when, and what were the questions?” So evolution is a common denominator for everything. And we started with these questions. Given a software project, it could be any project. I mean, the world runs on software, so somebody has to develop, put some code in there. When did it start? Who started it? Today, we might be reaping benefits of multiple softwares, but it’s kind of an academic as well as interesting project for DevOps as well. Say, suppose you have a 5GB worth of software. How did it grow this big? It didn’t happen in a day, right? So who introduced bugs and who fixed those? And a lot of times we say, “Hey, this software is complex, you have [inaudible] and other complexity metrics [inaudible].” This is your complexity - cyclomatic complexity, this is your cognitive complexity. How did it grow? Because as engineering managers, people don’t want to have complexity in their code base. How does it grow? How many actively [inaudible] at different points of time in a project? And how the functionality grew? What is the technosocial interaction between the developers, between those who write the code, those who review the code? Is there a way to do all these things?

Hari Prasad Sudharshan: 00:10:53.651 And we used to get - and then we also told everybody, “Hey, are you making real progress? Are you changing existing functionality again and again?” Assuming that you have a module and you didn’t write the module really well upfront. So what would you do? You keep on rewriting it. Can we get a metrics which says, “Hey, you may want to have a look at this.” And all of these are today whatever we have in engineering from my perspective, we look at all the project at 15,000 or 30,000 foot above the ground. So what really happens is something like - as we discussed, the theme of the meeting is not only evolution - it’s a blind spot. It’s a blind spot to everybody out there. So we developed an architecture to understand the evolution, okay? We required a software framework to understand, hey, softwares evolve. And this is our architecture. There you see [inaudible]. And the second one is the software solution - software evolution app. The software evolution app taps the data from the version control systems of software development. And the collaboration app taps the information from whatever code collaboration or code staging tools that you guys use. It could be anything, right? Currently we have like one thread going on, which is gate plus [inaudible] combination. And we know the developer pushes - so since I have my DevOps background. So first we started with this Influx based one because I could map anything that happens with respect to time. Hey, this developer is pushing one line of code here, and the next day somebody comes and changes it. And when [inaudible] as a bug, which means they change that yesterday somebody pushed is the root cause for this bug.

Hari Prasad Sudharshan: 00:12:51.940 So we were able to do a root cause analysis based on this, and we used SQL for normal, regular storage and used graph for predominantly linking - okay. Say, suppose you have a team of 10. Who do developers go to, to get their code review, right? Who reviews this code? These are social interaction among your team and you need Facebook kind of graph or Google Circle kind of graph. So we also develop that kind of a graph using GraphDB, and Influx is our main database, wherein anything that happens first goes here, and then those which qualify to be useful, we put it in here, and the mundane regular data, we put it in this statement. And then we have the blind spot compute engine, which works by passing all these, and we have ML models, and we have a statistical computing engine. And on top, you see review - we are planning to give, we have not yet started. We have few items which I can show you today. You can give not only the management. Anybody from top to bottom was working on a particular project. What’s happening? Say, suppose there is a new developer coming and he is making a change in a certain file without understanding that dependency happens most of the times in open source. So this model will say, “Hey, whenever this file has been changed, there is another file which people change, and the correlation is 90 person. You may want to check.” This will never come with any YouTube training or - this happens at a very surface level. It doesn’t happen like in a classroom session. So the more granular your data is, the more recommendations that you can give to the developer. It’s not only recommendations that you can give to, hey, he may be your friend.

Hari Prasad Sudharshan: 00:14:45.724 So that kind of recommendation can be given to modules and software. “Hey, this Java file [inaudible] might be another Java file. You may want to check.” And then while doing it, there is a possibility that we have identified that you may be able to write better code without introducing bugs to a certain degree. Okay? So I’d like to show you the thing that we have done. So all of you of - software evolution. Let’s do a simple overview. We took this project, Android OS. Android is by far very well-known open source software, so we picked up that and we picked up one of the major components that is the basic platform library code. So sorry, it is library core. And the analysis yielded some interesting insights. I’ll take you through those. It started in 2007, and it has more than a million comments, so that’s how when you use your Android phone today, one of the core reports of Android has close to thousand developers contributing to it. That’s why we get it for free, at least, the software. And the files commenter, as I told you, the correlation analysis with bugs and the dependency management. The files modified pattern, that’s what I’m going to show today with the little time that we have. So people from software background know that nothing happens out of the air. Somebody has to write a piece of [inaudible], right? And how many lines do they usually comment? And what is the complexity of those lines? And when they comment, do they comment one file or two file? This information are not very well known, right? So we will see those, and this is our phase one. We have completed our phase one. Our phase two is we want to tell a huge story to the world. “Hey, how did the phone that you use, which runs on Android, happen?”

Hari Prasad Sudharshan: 00:16:47.231 I mean, it’s kind of, I would say, an interesting goal, because if we can do this for this repo, I’m sure everybody can do it for their own repos because we’ve written it in the most generic manner. And I’m going to the next slide. So this is the crux. I’m going to show you some data. So looking at this data, each point here is a comment. The developer comments to the code base. Okay? I have not used this area or scale just because the points are more - as I told you, there are [inaudible] start. It started in 2007. On the Y axis, you see the number of files modified or comment. Whenever a developer’s pushing code, how many files does he comment together? As humans, we have a cognitive load, right? So as you increase the number of factors to think, you are bound to make mistakes. So this in itself may not be very useful. This is raw data. I will show you what we have done. So same data here. Just that the third dimension is the heat map that you see. It is just our interpretation. People can interpret the way they want. One being like, “Hey, I modified one file,” because it is generally believed that small incremental comments produce less [inaudible]. If you push more lines of code or integrate more files and comment it together, there’s a huge burden to do integration test and further so on. So it’s always better to push, okay?

Hari Prasad Sudharshan: 00:18:42.720 So if we ask any engineering manager, “Hey, how do you think your integration tests are going?” So integration tests would be very simple and faster if you have simple comments going every now and then, and you have a huge Big Bang integration test which can test everything. So talk to any engineering manager, they would say, “Hey, I want the number of file comments, files per comment [inaudible].” How many files are you commenting - I mean, how many files do you change per comment? People do not have answer because that’s the blind spot that we have [inaudible]. So if you have any questions, we will take it after this. So I just made it a little bit more readable. So this is same data. What you see on the top and the bottom are both the same data. And now we have a comparison here. So how did it start? Well, it started more than a decade [inaudible]. And how is it now? How people are commenting today? If you see, it’s almost same. And the number of developers have definitely increased. When it started, it had fewer developers. I also have a developer count. So on the top, what you see here is what’s happening. This is June of this year, and this is until November. How people are still commenting. You see here, what we analyze, we did a correlation analysis and found that initially, it looks like this project was already done. It did not start in 2007. It was already done. It was in some other place managed by some people. We don’t know the history, but if we ask this question, you might get answers. So what they did was they had a huge bunch of already developed software they wanted to integrate. So initially when the project started -

Hari Prasad Sudharshan: 00:20:40.947 So these numbers, I just wanted to reduce the scale to 512 so that you could see it well. You wouldn’t believe there were 7,000 files or 6,000 files put together. It is not humanly possible to make changes to 6,000 files and comment. Right? So you pick some component and bring it here to your software codebase. So those are the kind of activities which has increased the number of files for comments. So I just took this. So this 512 that you see here are - there are more. These are not changed by humans. These are done in someplace where this green has its own history. Okay? This 512 files could be developed by three, four, five, six developers over a period of time, and we could also backtrace this. If you know, okay, where did this come from, you could backtrace that. So whatever you enjoy today, it’s not just Android studying as work. It could be 20, 25 years, which has happened over a very large period of time, which involves a lot of developers, and it’s a huge story to tell. Right? So just a comparison. You see that there are a lot of - not much red lines. Here, the red scale is quite less. It is 20 here. It is almost compatible, maybe because the depth kind of different than these two scales. So when we did correlation analysis, people are still commenting with the same amount. The rate of comment or rate of delivery of functionality is almost the same, but the number of developers have increased and the complexity of the product has skyrocketed unlike anything. It is on the roof. So the business is always interested in delivering software. It’s the engineering side which has to worry about complexity and other things. “Hey, how complex is my code base for change?” If you don’t know the answer, then it is not going to solve you over a long period of time.

Hari Prasad Sudharshan: 00:22:37.598 Evolution is important. Why? Simply because you will know, “Hey, will I be able to continue at this pace? Or should I take a break and fix something so that I can go the next 1,000 miles with more vigor?” So that’s the questions that we wanted to answer. And this is the collaboration metrics, so. You develop code, and there is also something like, “Hey, you have developed and it’s in the staging area where people are waiting for somebody else to approve it. Could be a rubber stamp or people could do active review. And how fast is it happening?” So on the top, you see, the blue lines are the new comments that are coming in. People are actively pushing, “Hey, I’m doing this, I’m doing this.” And then on the green, you see, which is always lesser than the blue, which says, “This is approved. This is not.” I mean, for people who are not from software background, like everything, there is an approval process in software. It’s just don’t like, “Okay, I have a good code.” It doesn’t go just like that, right? So there is somebody who is writing your code. And how fast are they writing? You can also say that if the writing is not happening as fast as the incoming merges, then you have a bottleneck to address. So these are metrics that we found very fascinating and this helped us. This helped us big time. To make sure, if the green is always lower than the blue, then I have to go and talk to my reviewers or entries. Tell somebody, “Hey, you just review because that is increasing velocity, our speed.” And bottom, you see, this is like a very [inaudible] addition and deletion. The number is the of lines that is added to the report, as I told you. Today, it is like 10 GB of code. How did it become 10GB? Somebody must be doing it one line, two line, for a very long period of time, right? This is the last six month - over the last six months, if you see, the code has been - somebody has deleted - this is like close to 65 million lines of code and people are adding code as well, the max being 5,000 or 6,000.

Hari Prasad Sudharshan: 00:24:40.862 So people are actively adding code. Somebody is actually refactoring. So then the refactor [inaudible] functionality time, because when people take down some code, it means they have in the past worked up something so that they now know, “Hey, to do this, I now can reduce this. So I already have this functionality going up.” So it is like - I mean, Google follows the - they have this flag, which will say once the functionality is ready, then they can slowly deprecate things, and things are getting deprecated for refactoring or the functionality becomes obsolete. So you can tell all those stories really well. So to conclude, I would like to say, to do all this [inaudible] everything that you saw and we are planning to do and have done are all based on time. Imagine if we did not have this or we did not [inaudible] this database. Then I was actually trying to change SQL into Time Series and - yes, you can do it technically, but not with this much ease. It just removed our - I mean, I don’t need to worry about it anymore. That’s the one thing that I could say. And thank you Influx and team. They’re also very supportive on the Open-Source site. And Influx 2.0 is, I feel - okay, I’ll finish with this. We use Python for statistical analysis, but Flux has more functionality, so now we think, “Hey, if I have to do some computation on the [inaudible].” It’s not replacing Python by any standards, but if it happens, if I have to do Python, then I have to read from the database. And then it’s costly, right? I have to compute, and then I have to push it. So I could do that with Flux. So Flux has improved our main median mode and all these basic things, statistical things. So looking forward to more changes in Flux. Hey, thank you, Caitlin, for this [inaudible].

Caitlin Croft: 00:26:56.231 Awesome. Thank you, Hari. That was great.

Hari Prasad Sudharshan: 00:26:58.963 Thank you.

Caitlin Croft: 00:27:00.161 So if anyone has any questions, please feel free to post them in the Q&A. So Hari, you talked about searching for InfluxDB? Did you try any other Time Series solutions before using InfluxDB?

Hari Prasad Sudharshan: 00:27:18.873 We [inaudible] and we felt I could write - I mean, during initial times, it was pretty easy for us to get that up and running. And we did not - the pushed model is something that we were looking for, so it was quite [inaudible] access between the two, and we zeroed in on Influx.

Caitlin Croft: 00:27:47.982 Yeah, and how long did it take you to get familiar with Flux. I know a lot of people take a little while to get used to the new syntax.

Hari Prasad Sudharshan: 00:27:57.615 Okay. So it was initially a challenge. And we moved to Flux because [inaudible]. He thought that if you could program the [inaudible] as code, anything as code becomes easier because you can program. And it was in our motivation to see if you can manipulate database with kind of any [inaudible]. It depends on you. One of my bright engineer [inaudible] a week, and she did most of [inaudible]. Her name is [inaudible]. So she couldn’t join us today. I’ve shared this discussion with her and she picked it up really well.

Caitlin Croft: 00:28:50.428 Do you have any - do you have any tips as far as if someone’s trying to learn Flux for the first time?

Hari Prasad Sudharshan: 00:28:57.608 Okay. So the documentation. The current documentation is awesome. Okay? So I see a lot of changes in the documentation. Flux has a good documentation and [inaudible] initially fail, and I don’t have many tips. We just follow the documentation. And I think it depends on what you’re trying to do. If you are familiar with Influx, and then - see, what happened was we were familiar with the [inaudible]. The query builder helped us a lot. And it’s kind of sometimes [inaudible], sometimes it auto populates. So yes, the [inaudible] is one that I would suggest people use. It’s useful.

Caitlin Croft: 00:29:46.122 Yeah, I think that’s a really good tip. I know that a lot of people really rely on our documentation because there’s so much you can do with our platform and it’s just a matter of figuring out how.

Hari Prasad Sudharshan: 00:29:58.308 Absolutely.

Caitlin Croft: 00:30:01.485 And had you been manipulating a lot of Time Series data before this project? It sounds like you were somewhat familiar with the intricacies of Time Series data before.

Hari Prasad Sudharshan: 00:30:16.221 Yes. See, I’m fascinated with data, but for Time series, this is my first time. I mean, it just opened up new - I would say, see, before that, to get X axis as Time Series, I did not think [inaudible] at all. I mean, I knew something like this existed, and it opens up huge - it opens up your mind, to put it frankly. You can connect different points now. So I haven’t used it before Influx. This is my first [inaudible] activity, I would say.

Caitlin Croft: 00:30:48.803 And are you hoping to use InfluxDB at work or with other projects?

Hari Prasad Sudharshan: 00:30:53.773 Yes. So whenever I see - okay. I’ve seen Influx fixing everywhere because as I told you, anything you came up with - anything that we do, we came up with time, right? So I don’t see any place where you couldn’t fix it, for my hobby, weekend or official. I use it everywhere. I also tell people to use it because it’s quite simple to get on board. So you have the predefined - how do you write is already pre-defined - okay. One more thing is with 2.0, you have the goody UI, even for the on-prem, you have a good UI. And not only for cloud, right? So that is a major takeaway, Caitlin. You can write how much ever you want. Nobody is restricting.

Caitlin Croft: 00:31:46.709 Yeah, that’s the thing with Time Series data, right? You’re going to start off with a ton of data, so you need to be able to handle it all. So we have a question here. “My company is running InfluxDB 1.X. Can we use Flux?” So the short answer is yes. If you’re using InfluxDB 1.8 or anything newer, then you can use Flux. But I would definitely recommend checking out the latest InfluxDB, check out 2.0, check out our new - check out our updated cloud offering. There’s just a lot of features and it’s just nice that it’s hosted in the cloud, so it makes it super easy for you. It doesn’t seem like there’s any other questions as of right now. We’ll just stay on here for just a little bit more. Thank you, Hari, for presenting that. I think it’s great. I think it’s very cool to see how you’re using InfluxDB. And you’re using all of our products. You’re using InfluxDB, Flux, our templates. So really great job.

Hari Prasad Sudharshan: 00:32:51.499 Thanks. Thank you, thank you. Thank you for Influx and the community. Thank you very much.

Caitlin Croft: 00:32:57.189 Yeah, I love the InfluxDB community. Last week I was in Vegas at re:Invent, but it was so much fun getting to see users in person again. It was pretty exciting to get to meet you guys, so definitely keep it up. Love seeing all these projects. If anyone is doing something cool with InfluxDB and they want to share it with the community, please feel free to reach out. I’m always looking for new stories to tell, so.

Hari Prasad Sudharshan: 00:33:24.833 Definitely.

Caitlin Croft: 00:33:26.770 Thank you so much, Hari Prasad. I think you did a fantastic job. If anyone has any more - oh, okay. So someone has a question, and Hari, you might be able to give some feedback. “How hard is it to upgrade from 1.x to 2.x, which supports Flux. My company is using 1.x, and I’m afraid it’s hard to convince the leaders to upgrade.” So I will say this. It’s actually pretty easy. We’ve had a lot of people upgrade. We definitely see it as an upgrade versus a migration, which I know migration is definitely a scary word for people. I would definitely check out our documentation, and we also have webinars on the content written about upgrading to 2.X. Hari, did you start off with 1.X or were you always using 2?

Hari Prasad Sudharshan: 00:34:21.740 I started with 1.6, moved to 1.8, and we recently moved to 2.x.

Caitlin Croft: 00:34:27.661 So how was it moving to 2.X for you?

Hari Prasad Sudharshan: 00:34:31.673 [inaudible] contacted a few people in the Influx community and they helped us out. I mean, we did it when it was like 2.0.1, it was very early stages. That’s when we did it. And now I think things should be resolved. The dust should have settled by now.

Caitlin Croft: 00:34:57.279 Awesome. Well, Gong, if you have any more issues, definitely reach out. I know our community is always there to help out, and I’m also happy to connect you with Influxers who can help out with your upgrade as well. Well, thank you everyone for joining today’s webinar. It was really great having you all here. Once again, it has been recorded and will be made available for replay later today. Thank you, Hari Prasad.

Hari Prasad Sudharshan: 00:35:27.506 Thank you. Bye.

[/et_pb_toggle]