In this episode I speak with Martin Mao of Chronosphere about the massive scale monitoring platform. I also cover fixing the internet, fixing air conditioning, the problem with calculators, switching from macOS to Linux, the death of Lil Bub and more!
SPEAKER: M2 Welcome to the weekly squeak your weekly geeky squeak with me Chris Chinchilla. Now as you probably already hear a bit of cold or getting over a cold to be precise you have to put up with that sound of my voice. SPEAKER: M3 But anyway I have an interesting show. I’ve got my links for you and then I have an interview with Martin Mel of Crossfire ex Uber and other companies about their new I guess a big data streaming analytics platform built on the open source project M3. That’s coming up later. Let’s start with my favorite to the week. SPEAKER: M2 Beginning with an article from John fingers on Engadget. Always my favorite segment of the show in fact maybe I should actually just make a segment because it seems to be a topic that I bring up quite a lot. Oh technology. It is a couple of days ago. It was the 25th anniversary of PlayStation. The original Playstation was the Playstation still exists but the original one that’s actually also quite amazing if you think it will keep up with consoles very much but I guess it’s still up to PlayStation 4 maybe 5 is coming soon. That’s five models in 25 years one every five years which is pretty slow pace of release actually. SPEAKER: M3 Which is also kind of interesting in its own right. But anyway I digress. This is post dissecting the original PlayStation I acquired one at one point actually. And I guess the first thing that strikes you in the photo is how big it all is. You know there’s a lot of surface mount technology and devices these days it’s got a C D drive that you don’t think you have anymore in modern consoles. Again I’m not so sure it’s got huge resistors and capacitors and just a lot of space. I mean I don’t know if you’d want to service a Playstation but I guess you could and that’s something that’s harder to do these days with any modern device really. So yeah if you remember the Playstation I do it was pretty amazing and groundbreaking at the time. And now I could probably emulate it on my iPhone quite easily. I guess I don’t have an iPhone video to me and it look a lot better. But it was groundbreaking at the time and yeah have a little trip down memory lane and have a look inside literally the PlayStation if you so desire next to an article on Gen from medium by Microsoft. This is the story of big calculator how Texas Instruments monopolized math class or math scores have you prefer to say that I found this interesting I guess what I read this I thought this was going to be historical and it would fit into my historical corner but no Texas Instruments Graphing Calculators is I think I think I had one when I studied Mass. I can’t quite remember. I actually stopped studying maths fairly early on. So I don’t entirely remember if I had one or not or maybe my dad had one but that was quite some time ago. I guess I just assumed that things like graphing calculators had been replaced by apps and yes actually they have. But of course if you allow children to use phones and tablets then there’s all sorts of other distractions. So calculators graphing calculators are still mostly preferred in many schools especially in America. I guess where this article is based I don’t know about the rest of the world but despite them being pretty old technology they haven’t really changed very much in quite some time. Definitely the lifespan of some of the children that are using them. They’re very expensive. And if you lose one you have to get a new one. And in certain schools and certain regions kids just can’t afford them. So the schools or quite often which is even crazier the teachers have to buy them for their students so they can do the work. I mean we’re talking about very complicated here that you can’t do in your head or on a normal calculator was you could on a computer. But then yeah I don’t I I feel like there’s better solutions to this than than having one use very expensive device and maybe locked down on devices Chromebook or something like that that you could just restrict access to. I feel like is a better option. But sometimes inertia and lethargy just take hold in these sorts of situations. But if you’re fascinated to hear how you know a 20 something year old device is still. A common requirement Well I mean of course there’s pencils and pens and things like that. But anyway a technology device. Well it’s also pens and pencils. Anyway I digress. I think you get my point even if I’m not making it very well. If you like fascinated to hear how such an old and unthreatening device is still being used by children today and possibly many of you who are parents also used it. Then read on it it’s fairly long. It’s amazing to think that it’s possible to write so much about the Texas Instruments calculator. But it is and it’s quite interesting. So enjoy it. Now this is not a new post. It’s another medium post from woot I’m guessing he’s Dutch about switching from makeovers to Linux and his man he posts like this but I bring it up because I have been experimenting with this. Again one of the people I work for I discovered I was entitled to a work laptop so I thought here’s a good opportunity to try it again. After hearing so many people say to all Linux just works now just Linux is much better than it used to be. You could do video editing. Many of the things that my checklist of necessities to switch from Mac OS that always held me back on Linux. People say it’s better now it’s better. And after reading a couple of articles like this it’s become the kind of new cool thing to do switch from magazines to looks for certain sections of the computer using population. I thought I’d give another try. I’m not going to go into very much detail here because I’ve actually written quite a long article which I will publish on these very soon. I have a few extra things to add to it from the past couple of days and which will detail my experiences. Shall we say I’ll give you a little hint that it was all smooth sailing. But yeah I think I’ll leave it there. Maybe I’ll save a further discussion for next week. But if you want to do some preliminary reading around the subject then have a rate of this post and he has a couple of follow up posts as well. I especially found the section on alternative software quite useful although interestingly a lot of the software is not necessarily fully open source which I found interesting the concept of having proprietary software Linux has always been a strange one to me. It’s fine but it feels like the section of people who’d want to use that would be more. Would it make it worth it. But anyway that’s for a bigger discussion. But you have a read of this and then I’ll be back with my take next week or in the next couple of weeks and you can hear about my experiences directly on the subject of change and switching platforms. Tim Berners Lee is back in everyone away but he’s back with a new campaign. His global plan to save the web this was his contract for the web. It’s actually being announced over the past few weeks. He’s trying to encourage digital players to turn the Web back to I guess what he always wanted it to be. Although interestingly little bit of history here that not many people are necessarily aware of. I recently just finished reading the innovators by Walter Isaacson a very good book actually I really enjoyed it. I kind of could stop reading it. And in the section that covers Tim Mosley and the world wide web. Tim Berners Lee was actually generally in favor of keeping the web fairly restrictive to academics there so it’s interesting that he now a day people change its mind. He now kind of flies this flag for openness when it actually isn’t necessarily what he intended. And the browser manufacturers from the time like a mosaic for example were actually the people who pushed him and the World Wide Web to be more open. But anyway that was then this is now and I’m going to phrase this the contract for the web requires endorsing governments companies and individuals to make concrete commitments to protect the web from abuse. And in short it benefits humanity. It’s all very noble. But what does it mean when some of the people who have signed up for it are some of the companies that you could accuse of breaking it in the first place. So I’m not 100 percent sure what this is all really going to mean. I have been trying to line up an interview actually with Tim mentally for some time since Web Summit a couple of years ago. I don’t know if I’ll ever happen but I’ll keep trying and maybe we can dig into some more details. But in the meantime have a read of this article on The Guardian and have a dig around the contract for the web web site. And yeah I’d love to hear your thoughts. What is it. Is it is it binding. Is it just grandiose grandiose statements. What do you think. You can let me know. Chris Gillette dot com slash contact and another article on redesigning redesigning things that have not changed in a long time. This is from CNN from Jacobabad Briscoe and it covers Yeah redesigning air conditioners. Air conditioners have not changed in a very long time. If the world is going to get hotter we might need them. But ironically they are often the cause of a lot of climate change. It’s actually a statistic here let me just find it 10 percent which is fairly high. It’s not the highest but it’s fairly high. And if you consider that not all the world really uses air conditioning then that starts to to show it would probably a lot higher in certain countries. So how do we make it more efficient if we’re going to need it. Now recently there is a new coalition from the Indian government in place that definitely needs it. And the Rocky Mountains Institute Rocky Mountain does not sound like a place that needs air conditioning but still they are offering a one million competition to improve air conditioning. And this article details some of the products and I am super fascinated to see how this goes. I have often looked at air conditioning and think thought yeah how can we make this better. We all know it’s terrible the environment and very expensive and it blows grids and we lived in Melbourne in summer often peak summer when there’s lots of air conditioning. You have brownouts the power go out all the time so you’d be a problem in many many nights. So I look forward to seeing the results of this and maybe in the near future even sort of speaking to some of the people behind similar projects and seeing what they’ve come up with and then how long it will take for them to become commercially viable. And on the streets in our houses and keeping his cool what’s the world is hot but not heating the world up even more. And finally very sad little bub the little kitty with the sticky out tongue made everybody very happy has died only eight but I think she had health issues so I guess that explains a lot. Very sad. Internet means everywhere. Cried out. And the interesting thing here is that a lot of the famous meme cats are all dead now keyboard cat is dead. Grumpy Cat. I think he’s dead. Little Bob is dead. About the only internet cat I really know. I said Hey they’re all gone. So who replaced them. We will see. Maybe it could be our cat could be onto a fortune there anyway. Spare a thought a little bob. They’ll face her little face or die like a female name doesn’t mean anything. Sedative to get your tongue for one minute silence. I hope you enjoyed my links this week. And now I have an interview with Martin now of currency enjoy. SPEAKER: M4 So our journey began at Uber four years ago and what we realized that Uber as we were going through the migration in adopting Cuban it is and a cloud native architecture and adopting like my free services and running everything on containers. SPEAKER: M8 We realize that none of the tooling that existed there could handle the sheer volume of monitoring data being produced. So none of the tooling could even store all of the data being produced let alone sort of do anything useful with it. And as we were going through this with Uber we sort of looked around the market couldn’t find anything. We ended up building our own solution which we actually built in open source and that solution is called M3. And during our time at Uber we actually managed to scale M3 up to one of the largest monitoring systems in the world today. SPEAKER: M4 And then earlier this year we have decided to leave Uber and build Kronos sphere the company around the open source M3 Technology to sort of bring all the benefits that we sort of extracted with M3 and bring that to sort of enterprises all around the world. SPEAKER: M5 So I mean there’s a lot of monitoring Stacks out there that claim to be to do the same thing. So what’s the problems they have. What’s the limits they have. SPEAKER: M6 Yeah it’s a great question. So there are definitely a lot of monitoring solutions out there that claim that they are enterprise grade. But when you look into it it’s generally a few things that they fall short at. SPEAKER: M4 The first is the scale and the amount of data that they could store. If you look at a lot of these solutions that are enterprise grade they store tens of millions of metric Times series. And what we found is as you get more complex technology stacks and as you sort of move to communities what ends up happening is you just generate a lot more time series than they used to. And if you look at something like three you can actually store tens of billions of times three. So it’s actually not even just an order of magnitude but it’s multiple orders of magnitude more in storage which is great. And then the two other things that that really do differentiate it is with a scalable solution that’s that’s that’s really nice but it needs to scale in a cost efficient way so it can’t just scale in a in a linear fashion in terms of cost. But we really made sure that as we scout up the system to storage we scale it up the costs sub linearly. So you’re not paying as much for the same amount of storage. And then finally the the big differentiator for enterprise is reliability. So if you look at a lot of these other sort of solutions that are out there they run on a single region in a single cloud provider. And because of that they can’t really provide high levels of reliability. And that’s one thing that sort of plagued our system at Uber many years ago. So what we ended up building especially with Kronos sphere is a solution that runs across multiple regions in multiple cloud providers so that you get the highest level of reliability for your monitor. SPEAKER: M5 Would it be fair to say that you’re starting to get into maybe the levels of technology and infrastructure that many might reserve for their main application more than just just just there they’re monitoring stacks for sure for sure. SPEAKER: M6 So yeah I think if you if you look at what we used the underlying technology for and what a lot of companies use it for today they’re monitoring not just all of their infrastructure and their applications but actually quickly use the tool to monitor that business as well. SPEAKER: M4 So so the use cases definitely sort of expand across the whole business. SPEAKER: M5 OK. So let’s let’s get into three first and kind of go up go up the layers. I mean I guess it’s open source so we can go into some detail on what what did you do to to make it so much more optimized for this large scale use case. SPEAKER: M6 Yeah it’s a great question. SPEAKER: M4 So when we first designed the first monitoring solution in at Uber before M3 we sort of used off the shelf open source technology for both our storage and our metric index as well. So we’re using technologies like Cassandra and elastic search and what we found is that those technologies are just not purely tailor made for the time series use case right. Cassandra is a key value store an elastic search is like a search engine. But what we ended up doing is beyond a certain level of scale when you need a certain level of performance and cost efficiency you really need to build everything from scratch. So the first thing we did with them three was to build a completely new time series database as the underlying storage engine and that is built completely from scratch. It’s not built on top of any other sort of existing storage technology. And then on top of that we sort of then layer on and ingestion pipeline that’s a highly reliable and can do things like metric aggregation industry infection and then also a scalable Query Engine well because you know storing billions of Time series is great but once you have a store and you want to be out a query billions of Time series so all of those pieces of technology we just found that at the scale that we were operating at and the performance that we required nothing out there really solved that solution. So we really have to do everything custom from the ground up for this one particular use case. SPEAKER: M5 And as far as I can tell you just focus on the ingestion and the storage and kind of hand off the the the other stuff to do other applications like Prometheus for example. SPEAKER: M6 Yeah. That’s that’s that’s sort of correct. So. So definitely sorry. SPEAKER: M4 So I definitely it does the ingestion the storage and the query engine for sure we in the open source solution we actually don’t provide any of the visualizations or the lighting or the tooling there in terms of Prometheus. We do plug and play with Prometheus so people who are running Prometheus today because you know Prometheus is great and it’s the thing that everyone gets started with but it doesn’t solve it’s not. It runs on a single binary so it doesn’t really solve the scalable high reliability nature of monitoring solution. SPEAKER: M8 So M3 in open source actually plays really nicely with Prometheus in the sense that it’s the long term storage for you. Prometheus points which is great. But yes in open source you definitely have to sort of piece it together and that’s what we do differently with chronic fear is that we didn’t just want people to have these things together we wanted to both post everything for everybody so that we could run the complex M3 stack but also provide the visualizations and the already tooling and analytics on top in all in one product. SPEAKER: M5 And how long did it take you to build. I mean I’ve known some companies that have been trying to build something like this for years. SPEAKER: M7 Yes it definitely took us he has to build it reject it open source April 2016 and we’re still actively chipping away at it in open source and adding features every day so it took more than three and a half years to get to where it is today. And that was developed by a team of 20 senior engineers I was heavy investment in the program. SPEAKER: M5 So let’s move up the stack to CROSSFIRE. 1 What is Kronos sphere beyond. I’m guessing M3 as a service. Like I’m guessing I’ve had more than that. What do you add on top for sure. SPEAKER: M6 So what kind of sphere does is first. It starts with M 3 and actually we add a whole bunch of preparatory features that are not there and all these features are tailored around the enterprise use case. SPEAKER: M4 So with open source M3 you can actually scale a single deployment of that up to billions of Time series. That is possible but when you look at an enterprise what you really want is sort of multi tenancy controllers that sort of map to your organizational structure. So you want teams to be able to sort of emit metrics that other teams cannot see. You want teams to have sort of resource allocation so that one does not step on the toes of another team and take the whole system down. And you also sort of want security and access controls as well. So all of those type of features are just not bundled with the open source and free solution today. And that’s the first thing that we we layer on top of that with the view. And then the layer above that is what I mentioned earlier with the visualizations be alerting engine and the analytics tooling so tooling that let you leverage the amount of data that you storing now all of that is also not available in the open source M3 solution you sort of have to piece that together yourself. So so that is something we provide. And then finally we run the whole thing we we manage the whole thing. But we run that across multiple cloud providers for you so that you know you get the highest level of reliability without having to worry about how how you’re going to sort of structure that and take care of all of the underlying structure for that. SPEAKER: M5 And I mean when we met a niche very briefly at the velocity one of your colleagues started to show me a visualization as much as possible with no visuals. Could you maybe just talk through how that works and what you what those visualizations would be useful for in comparison to I guess other types of visualizations that other people might have used. SPEAKER: M6 Yeah for sure. So first and foremost we provide sort of the bare bones essentials of a monitoring system where you can sort of create dashboards and graphs and start graphing your metric in your time series data and then we have an learning engine that lets you set a threshold based alerts on top of that data so you can be notified in that that alerting engine sort of ties into the standard notification engines like page duty and can send emails and things like that. SPEAKER: M4 So that’s the sort of basic solution that we have in that we can do but what we realized throughout time at Uber is that the amount of data actually grows at a much faster pace than the amount of engineers you hired to use this tool. So what we ended up having to do is to automate a lot of that. So if you look at our platform today while you can create custom dashboards what happens is when you deploy our agent and our agent runs on humanities as a DeMint sets there’s like a one click deployment it automatically discovers all the metrics and points to scrape and ingest metrics. And then once it’s starting to ingest them it actually automatically recognizes a lot of those metrics because a lot of the systems you’re monitoring are very well known. So if you’re monitoring like an elastic such cluster or in my secret instance it’s actually fairly well known what those metrics are going to look like already. And then based on that we actually go and generate you all of your your dashboards for those pieces of technology automatically. And also the alerts for those as well. So you know as soon as you log into the system once you’ve deployed the agent you have a lot of your dashboards regenerated and ready to go. So you don’t have to consider anything manually so you sort of get monitoring out of the box of free and then the final thing on top of that is with the alerting engine. SPEAKER: M8 We also have anomaly detection running in the background because the thing is it’s actually also it’s not just quite hard to define what your light should be. It’s actually very hard to define what the threshold should be because that threshold changes over time depending on the load of your system depending on the time of day depending on the time of year. You know like it’s a Black Friday for a lot of e-commerce stores. They get a lot more traffic than they would on a normal day. So what our anomaly detection sort of system does is automatically generate the alert thresholds for you. Based on we can week or year on year analysis so it can sort of predict what sort of expected traffic loads are like and it set the thresholds accordingly to that. So the whole system is sort of while it provides the basic during the week so you could do everything. So what we really want to do is to automate as much of that as possible and just get you a working experience straight out of the box. SPEAKER: M5 And I’m guessing with that automation you mention that sometimes teams are discovering metric in points that maybe they didn’t even realize were there. SPEAKER: M6 Fisher For sure yeah. No it’s definitely happens a lot. SPEAKER: M8 And you know we definitely give controllers the team such that you know not only can they control which sources they want to ingest the metrics from but also that each individual source they can actually control both the retention period and the resolution so you can imagine that you know a team has exposed some metrics that are heavily used for debugging and you might not want to keep those around for years because you’re going to be paying expensive storage costs for those so I will actually give a lot of controls on defining not only what gets adjusted but at what sort of level of detail do they get adjusted for how long you want to retain the data for example. SPEAKER: M5 So let’s let’s actually go to some of the the business side a little bit. The company as far as I can tell has not existed for very long. Yes. And yes you managed to raise 11 million Series A. Which is which is a fair bit for four especially for a tech focused company. Well I don’t I’m not going to ask you like how. Because firstly I know really what I talk about. And secondly you probably don’t want to tell me but I guess to to raise that sort of fund so early. I’m guessing. I mean first the technology is proven you’ve sort of already used it in a very large well-known company. But what were some of the use cases that the funders saw would be very valid for particular businesses especially that that made them kind of invest such a large sum at such an early stage. SPEAKER: M6 Yeah that’s a great question. SPEAKER: M8 So as you said you know we we had a sort of boosted start where we already built this the core underlying storage engine and everything in open source so we could sort of build a business on top of that. So it’s much faster for us to sort of get to market with the end to end product than if we had to build the whole thing from scratch. But a lot of the investors when we started talking I think they came to the realization that this migration to communities in cloud native is well under way for enterprises and they came to the realization that a lot of the existing tooling out there today isn’t really built for this sort of new generation of use cases. And you know talking to a lot of the existing customers of these other solutions they hear the pain points and those are the exact same points that we are sort of trying to tackle which is you know what are these solutions are very expensive as you get to large scale and they’re highly unreliable. So those are the sort of exact things that we are aiming towards and those are the sort of design principles that we had when we brought him three from the beginning. So there was definitely a lot of alignment there and I feel like the size of the funding was also because we’re just much closer to getting a product to market. It was it was less about building the product and finding the fit. We actually had a large pool of open source users even of our solution out there in the world that wanted a little bit more than what we just had in open source so it was it was less of a exploratory phase for us and more of a execution and just getting the product out to market. SPEAKER: M5 And I think that was the reasoning behind the large round of funding and without divulging who you may or may not be speaking to. What are some of the common use cases or business sectors so far that have shown interest. SPEAKER: M8 Yeah we’ve actually I I I’ve actually tried to do the analysis to see if there’s a particular sector or segment that that you know has heavily relied on this and actually it turns out it’s quite spread across all the sectors so you know without going into sort of particular names we see you know companies are in the finance sector we see companies in e-commerce we see companies in more traditional retail. We see companies definitely heavily in the tech sector as well so it’s actually like a monitoring and because what you can use the underlying solution for which is not just monitoring your infrastructure new applications and you know if you look at a lot of businesses these days there’s definitely a huge push in technology and they know that that is a huge part of most enterprises these days. But outside of monitoring that you can also use the tool to monitor your business. And so Uber was using it to actually monitor all of their products across all of the cities and that would actually go and notify you know operations people not just engineers but operations people on the ground to let them know when something is wrong with a particular product in a particular city. So I think if you look at those use cases that can be reapplied across pretty much all or segments of the market that’s what we’re really finding at least I was talking and also a paper on as well is that we sort of get companies from all walks of life leveraging this technology could you dig into a bit more of what you mean by monitor your business do you mean from an operations side. SPEAKER: M5 Okay. All right. Okay. SPEAKER: M6 Yeah. Because it’s mostly heavily on the operations side. SPEAKER: M8 We purposely do not because it’s still an engineering tool. So you know we purposely do not tell. Well we purposely told companies to not rely on it for you know strict accounting purposes and why not. Although you can definitely use it to get a real time snapshot view of things like that. And that was definitely a previous use case and the current use case but we just sort of tow companies to use it as a real time monitoring in a real time signal of things as opposed to like the source of truth for the actual revenue or you know sort of financial base use cases. SPEAKER: M5 And I guess across M3 and Chrono Sphere bearing in mind CROSSFIRE is pretty pretty new. What is the sort of rough. So it’s hard to tell with open source projects but what would you say the kind of rough installation base is at the moment. SPEAKER: M6 Yeah. Great question. SPEAKER: M8 So insulation basically we actually don’t have exact numbers on that we don’t have like a phone feature or anything like that but I think if you look at you know we’re gauging based on things like GitHub stars is like more than twenty four hundred of those and I think it just based on the sort of community meetings we have. Definitely I would say hundreds of customers or like hundreds of oh sorry not hundreds of customers hundreds of open source users out there of the technology but I think if you look at something like M3 it’s going to be less pervasive than something like Prometheus because it’s not the tool that you start off with right. It’s the tool that you need once you get beyond a certain level of scale. And I think you know in that particular segment of the market there’s definitely a lot higher pervasive use than across the broader sort of SMB market. So you know some stats that we know we have it’s running in production today like 15 of the Forbes 2000 for example that we know of. And then if you go onto the get hub page and look at some of the the the stargazers I guess the people who have started the project you’ll see that a lot of the companies that I work at are these large enterprises Bishop but we don’t have exact validation on the exact number of installs because we don’t really do anything like that we don’t really like when technology does things like that. And then in terms of the pay product you know where private beta right now and we are onboarding a handful we’ve already onboard board a handful of of these customers and we’ll be looking to sort of expand that at a much more aggressive rate early next year. SPEAKER: M5 It’s actually a little bit of attention to question but I was just clicking around the M three getting started guide and came across so aside from Judge is a small part of what I do because tech journalism is you know it’s not the most it’s fun but it’s not the most lucrative of jobs. They actually mostly do a lot of technical documentation work. I came across hosts in your own in the M3 dogs not written by you written by someone from open source by the looks of it. I’ve never seen before in open source projects and tips for writing documentation and actually a lot of them I would echo 16 tips that I mean I’m not saying you have kind of any input or opinion on this but is sort of usability and understand ability of software away something that’s been important to you and the co-founder and I guess other people involved with the project for sure for sure and in fact if you look at the documentation I mean we have documentation there but we actually don’t love it as much. SPEAKER: M6 We don’t think it a. SPEAKER: M8 It definitely needs to get to those other being a couple of reasons for that while we were inside Cuba will sort of mostly focus on building the thing. And while we had a lot of outside interest it’s been much easier for us to sort of hold community meetings and talk to companies face to face just because the underlying pieces as we were building in the years past were changing so quickly that the documentation would be outdated very quickly so they would go read documentation that’s not really reflective of the state of the world I guess. But now that we’ve got a Cronus view we definitely want to invest heavily in our team inside Cuba didn’t really have a dedicated Deb advocate writing documentation or anything like that. So now that we have fits and Cronus V We’ve actually just hired our first advocates because what we definitely want to put a lot more effort into into documentation and supporting community. Good. SPEAKER: M9 Good to hear this. Yeah. SPEAKER: M8 The documentation is a bit raw because I think there’s also been something like this again and not something like Prometheus we can sort of pick it up and get started. SPEAKER: M5 There’s like a large complex distribution business you’re right so we sort of assume an underlying level engine here but that’s a whole other conversation around that kind of thing. All right. Let’s wrap up with. Obviously things are pretty new. Pretty fresh so I guess everything is is in flux and in plan but sort of most significantly what’s on the roadmap for the next six months or so. SPEAKER: M6 Yeah. Great question. So I think for the next six months you know there’s there’s some features we want to build back into the open source and free community so you know even though we have we have the paid product and credit sphere and that’s part of our concentration we definitely want to upstream a lot of changes back into open source inventory. SPEAKER: M8 So there’s gonna be a lot of performance and sort of additional reliability and like things like backup and features like that that we want to push back upstream for the larger community. And then in terms of Kronos video we’re sort of looking at Kronos sphere as a tool that you want to use to solve a particular use case in this use case is to get alerted when something goes wrong and sort of get to a point we can figure out the root cause and remediate the issue. So if you look at our road map there you know we’ve just added tracing integration as a feature to our product a couple of weeks ago. And you know it’s not like we are trying to sort of be a one stop shop for heads of ability and offer like an end to end tracing product but what we see tracing come in is you know once you do get a loaded based on your metrics you want to find additional details and context on what’s going on and that’s what tracing is great for. So we sort of added that integration in a way that it’s deeply integrated so you use the same client to admit both metrics and you trace this through broke through open telemetry and then that’s actually stored and deeply linked in the dashboard to go straight from a metric data point to one of the traces that underpin that data point. So I think if you look at our road map we’re going to continue go down this path of like what else can we do to optimize that particular use case if you want to set up some monitoring get alerted and remediate your issue as soon as possible. So that’s what’s driving a lot of the feature development for a host of products. And then as I mentioned earlier we’re sort of up streaming a lot of changes back into the open source community and I think we’re getting to the end of the year now. SPEAKER: M5 You’ve just come out of Thanksgiving and stuff like that too. Are there any events on the horizon or even local meet ups and stuff where people can come and take a look at what you do. SPEAKER: M6 Yeah great question. SPEAKER: M9 So we’ve actually since I think the sort of conference season has just ended it was quite quite busy but yeah we started looking at our content schedule for next year already and we’ve we’ve we’ve identified quite a few So in Q1 you know we’ll be attending a fund them of uranium which yet it’s enough and then I should be there pretty much every year. So. That’s awesome. Yeah. No I actually really like you know there’s no like you know we’re not gonna get a bit there to get a fork and and just just have a meetup for people using him three and whatnot. SPEAKER: M8 So that’s one we love every year. We will be doing cube on EU towards the end of Q1 next year and then we’re also going to be partaking in it Deb upstate as well as those look like much more city local type of meet ups. So we’ll definitely probably be looking to do the one in New York which is in Q1 and then also I believe they are in Seattle later in the year as well. So where we have engineers and you know we have four offices spread around the world so wherever we have engineers we’ll be sort of trying to partake in those and some local Meetup somewhat out as well. SPEAKER: M3 That was my interview with Martin now of CROSSFIRE. Hope you enjoyed that next week is my final event for the year arbeit Devereaux con in London. If you are in that part of the world come say hi love to meet you and then I’m taking a bit of a break until the new year when I think I’m going to be at CBS and South by Southwest. So big trips coming up. Actually it’s not very long wait. Only next month isn’t it. But anyway it’s next year. Though it seems like such a long time away. In the meantime you can find some of my writing on the dot com slash writing and more things there. I am going to be releasing a few new things over the Christmas break for you to enjoy. Actually I’ve got some new articles in progress that nearly done as well. So more there. And write ups from the various conferences I have been at the past few weeks which were data natives. I actually had some quite interesting stuff there and it sort of diluted all diluted to do I need to digest it all and summarize it. I’m hoping to actually do an interview with the company’s founder too. I did a workshop on removing bias from artificial intelligence and machine learning datasets super fascinating which I will get into more soon as well. And then I went to 5G technology. Again we learnt more about 5G stuff. Then I got kind of sick. So it is a bit muted but I have some interviews and things there to get out to you as well. So more to come in the next few weeks before the end of the year. If you have enjoyed the show please write review share. Love to hear from you. And until next time. Thank you very much for listening.