All aboard: How Irish Rail built an intelligent railway monitoring platform with Grafana

When Irish Rail needed to modernize infrastructure monitoring across the national railway network, traditional consultancy approaches failed. Multiple million-euro, proprietary solutions were disqualified due to rigid dashboards, vendor lock-in, and systems that looked impressive in PowerPoint presentations but couldn't meet operational requirements.

In this session, learn how Irish Rail got on the right track by leveraging Grafana's open source plugin ecosystem, data source flexibility, and community support. Today, Irish Rail Intelligent Sensing (IRIS)—a production-grade IIoT platform powered by Grafana, MQTT Unified Namespace, and TimescaleDB—monitors everything from track-side pumps and bridge sensors to passenger platforms and critical infrastructure across Ireland's 2,400 kilometers of railways.

The team walks through how it scaled from a single sensor proof-of-concept to ISA-95 compliant enterprise monitoring integrated with SAP and ServiceNow. With IRIS, Mean Time to Notification (MTTN) for safety-critical events has decreased, preventing bridge strikes and enabling predictive maintenance.

One unexpected result: Grafana didn't just replace monitoring dashboards. It became the foundation for Irish Rail's AI-powered operational intelligence platform, positioning a 180-year-old railway operator alongside digitally mature European leaders.

This is a story about choosing adaptability over vendor promises, open source collaboration over proprietary lock-in, and how the right tooling empowers engineers to build what consultants couldn't deliver.

Richard Donovan (00:00):

How many all just came for the AI like I did? Nobody, they're very polite. How many are using AI in actual production in their environments? Good, so that's quite a few. So we are gonna talk about something a lot older. Rail systems, which are quite old. But I think you'll also see it's kind of the future of maybe where a lot of people are going with AI, doing the development of the websites and the apps. Then our technology is spreading out into the real world. And this has been a challenge for Irish Rail, which we'll talk about, and it's probably where a lot of us will end up working. So I'm Richard Donovan. I'm a technology consultant that's been working with Irish Rail for four years now. And Seosaimh.

Seosaimh Ó Fátharta (00:44):

And I'm Seosaimh O Fatharta. I'm the IRIS Project Manager with Iarnród Éireann, and I've been in Irish Rail for a bit longer than Richard, about 20 years. And so I'm gonna talk about, at a glance, of who we are. So we're obviously the Irish National Railway company. Like all other railway companies, we were founded 180 years ago and we have almost two and a half thousand kilometers of track. This whole thing started off about a year and a half ago when I met Richard, and I was given a task for a remote monitoring project. I thought I knew what what remote monitoring was at the time, but it turns out what we've been doing for the last 20 years was very remote. We had loads of instrumentation and devices out in the field, but none of them were integrated. They were effectively just very remote, out there, not talking to anything except for themselves.

(01:31):

And we would physically go out with us with a smart card or something and download it and bring it back to the office. And then about five years ago, I suppose, IoT came along. And it was a game changer for us. We were working with a lot of surveying companies like Trimble and Hexagon, and a lot of smaller, tiny companies. But they all had the same thing in common. They all had their own dashboards. They all had their own systems. They all had they're own UIs. And it was great. We were finally getting visibility on information in real time, analyzing patterns. Like, I was in heaven. I was like, I don't need to go to site anymore. There's a battery powering the unit. It's sending the information up, and I've got instant visibility on it. And that's when the problem started. Basically, we were getting more and more devices. All of the devices were sending information up.

(02:26):

There was a few slight technical problems. Like, the lads on the ground couldn't remember their passwords, they'd lose their passwords, they'd go on holidays. The wrong mobile number would be input in the system. And I had no control how to manage the system.

(02:41):

As you can see, it was getting very, very frustrated on how to manage that kind of base that was growing out of control, because we were going from like having four or five devices to.

Richard Donovan (02:53):

Yeah, so we had over like, 50 plus vendors, everybody saying that they had the dashboard, their platform that we should all integrate onto.

(03:03):

They had no desire to actually engage with users across the board. And this is safety critical. So our time to under understand what the problem is, is key as you, again, as you'll see, as the businesses move out into the real world. And we're dealing with real data and real world scenarios. So this was very frustrating for Seosaimh, and we didn't really have a solution. So as any big state owned organization that Irish rail is, they went through a procurement process. Lots of documents were written, lots of requirements were gathered, lots of people engaged with huge lists of IoT devices were gathered and that we should all work through. And this went on for months,

Seosaimh Ó Fátharta (03:44):

Months and months and months. And I remember it well. We were sitting down with all the procurement executives, and they're going through like, the process of, you know, what kind of instrumentation have we got, kind of do this, kind of do that. And from my experience of 20 years, I knew what the machines could do. Like, I knew the machines inside out. I could repair them, I could fix them, I could move them around. But what I couldn't get is the finished product. So unlike a lot of solutions, so maybe some of them, I jumped right till the end. I said, this is what we want, Richard. How the hell do we get here? And we just bypassed everybody in the middle. So we kinda reversed-

Richard Donovan (04:17):

Don't tell 'em that. Don't tell 'em that. We followed the process.

Seosaimh Ó Fátharta (04:20):

So I suppose this is where the round applause happens. Whereas Richard just showed me Grafana for the first time and I'm like, what is that? And it was all black, and I was like, couldn't really understand it. And he was, this is what we want.

Richard Donovan (04:31):

He still has a problem with the black, actually.

Seosaimh Ó Fátharta (04:33):

Yeah, yeah. So you know, the black theme in the background in Grafana? You'll see later on in my slides, I use white, but I've got a confession to make after this event. I'm going back to the office tomorrow. I'm gonna change 'em all to black because all the guys up in the science fair. They-

Richard Donovan (04:50):

A conversion.

Seosaimh Ó Fátharta (04:53):

Their slides are brilliant. Their dashboards are class, they jump out oat me. I really feel it. So yeah, I'm a convert. Thanks.

Richard Donovan (05:02):

And so this is part of this whole process. Again, this is probably not the most technical presentation, but this is an opportunity for you working with Grafana. It is so flexible working with end users, especially if the end users are not also DevOps or SREs. You're working with real people, real engineers who are used to doing physical things, and you turn up with what you think is a super sexy Grafana dashboard. And often these guys don't like it, don't like the look of it. I've worked on projects where we built beautiful dashboards and we showed it to the customer, and they said "We hate it." And we said, "Well, what do you want?" He says, "I want a big green dot over there in the corner, 'cause we never look at this, and if anything goes wrong, it turns red. And if anything happens, then we'll call you."

(05:43):

Like, this is the early stages of engaging with a customer and actually finding out what they want. And sometimes less is more. So we went through this whole procurement process that was going absolutely nowhere. We had all these vendors trying to sell us super expensive systems all the way up to SAP that were very bespoke. That would take six months to produce anything sort of demo. And if we wanted to make a change, that would be another six months. And obviously, that's not the world I came from, and I didn't think this was good for the solution. So I started introducing this more adaptability, basically from my laptop, I started showing Seosaimh a Grafana dashboard with mock data from a laptop saying, what do you think of this? Did this start looking like what you might want? And he said, "I like this bit,

(06:27):

I don't like these bits." And we started iterating very quickly with our chief subject matter expert, and then started expanding it out to other engineers and other stakeholders across the organizations. We had no data, but we had...

Seosaimh Ó Fátharta (06:41):

We had plenty of data. It just wasn't integrated.

Richard Donovan (06:43):

We had no access to any data,

Seosaimh Ó Fátharta (06:44):

Had no access.

Richard Donovan (06:45):

But we had a concept now that the engineers were buying into. They liked the look of it, they could understand it easily. And if we could get them more of this, they would actually buy into it. So how do we get them more of this? So from a laptop, we start off with an MQTT broker, and start searching for existing data providers across the organization.

Seosaimh Ó Fátharta (07:06):

So I'll

(07:07):

jump in there 'cause I remember the day. Like I said, we're working with multimillion dollar companies and small, tiny companies with one person in his bedroom delivering a sensor device. It was a lot easier to work with one individual. He was, I remember him well. He was a small operator with a little weather station, and he was just providing us with temperatures and moisture content. And he went on board straight way. He goes, look, all you need is an API. I came back to tell Richard, and he goes, I didn't know you knew what an API was. I still don't, but I got it. And Richard is like, perfect. We have our data. We took that, we ran with it, and with what, two days? We had it up. And the managers just couldn't believe what we had done. Not only did we put it up, but we used the alerting system on it. And it was in the middle of the summer, and it was a really, with rail and with temperature, we can't run, nobody can run trains when the temperature gets too hot in a CWR, continues welded rail, because you have a chance of buckling.

(08:12):

So you get a hot summer or a hot day, and sometime in the mid-afternoon, which it-

Richard Donovan (08:17):

Ironically, it does happen in Ireland.

Seosaimh Ó Fátharta (08:20):

Oh yeah. And it happened in, I think it was June or July. But anyways, we had built a system not knowing that this weather was coming in, and we had an alert system built in, and Richard asked me, give me some phone numbers just to trial this on. So I fired out a few phone numbers to some of our managers, and all of a sudden, there's phones ticking, getting alarms. What's this? It's like 44 degrees somewhere in Cork. And they're calling me. "Seosaimh, what are you sending us?" And I'm sending, that's the data. Unknown to us, my original supplier, Declan, his system wasn't working properly, and he wasn't sending any text messages even though we were paying for it. His system was like, what? 10 years old. And it was, it was our take.

Richard Donovan (09:01):

Or we would say it was elegant and simple. It was one little box and a SIM card and some sensors wired up to it.

Seosaimh Ó Fátharta (09:08):

Yeah. Once Richard bought it into the IoT world, it was going down that way as opposed to poor Richard not paying his call card for the-

Richard Donovan (09:18):

That was the key thing. So we started sending out alerts to the end users saying your tracks are getting hot. And they're going, how come the system of record is not sending us any messages? And we found out the guy was running a prepaid SIM card, and he had forgotten to top it up. So it went from why are you guys spamming us with bogus data with, hey, you guys are now the belt and braces because we're safety critical. We're not frontline safety yet, but you guys are now the belt and braces that we can see what's going on. You can gather the data and you can start sending out alarms and alerts that really provide actionable intelligence to the people. Right? And that was one sensor. And they're used to dealing with one sensor. But now we were starting to get that traction where more people are coming to us saying can you work with this sensor? And again, with the platform being so flexible, we could start adding more and more sensors onto the platform.

(10:07):

So we had now some user engagement being, again, a sort of risk averse organizations, big organizations. Only when they start to see a real platform turn up, which was at zero cost at this stage, would managers start to get involved, because, you know, it's good for their reputation to have something, and if there's something concrete there, it makes a big difference. And Seosaimh came up with a genius idea to brand it.

Seosaimh Ó Fátharta (10:33):

Yeah, like people have ideas all the time, but trying to get it out there and trying to get 'em to believe in it, you only get one chance at it. So we had a demo day, or a continuous improvement day in somewhere like this. Not as big as this, obviously. This is massive. And I told Richard, we just need to get it out there. It needs to be flashy, it needs to be cool, it needs to have a bloody name as opposed to Grafana. So we're calling Grafana, we can't call it Grafana. We have to have our own name for it. So we came up with IRIS. You came up with IRIS. I came up with IRIS. So originally, it was the infrastructure, remote monitoring and inside systems that was very, very engineering, very civil engineer. We said, you know what? It's bigger than us. Let's call it something else. So we called it Irish Rails Intelligence Sensing.

(11:17):

And since then, from being aN infrastructure project, it's grown to the whole company, whereas we've got like BNF building and maintenance looking at it for elevators. We've got-

Richard Donovan (11:27):

Yep, cameras.

Seosaimh Ó Fátharta (11:28):

We've got cameras. Do you wanna tell them about that?

Richard Donovan (11:31):

No. Everybody, actually what-

Seosaimh Ó Fátharta (11:34):

What you can do with a camera is amazing.

Richard Donovan (11:36):

The cameras in the lifts they have turned off, because things that happen in the elevators that nobody wants to see, but that's another story. So-

Seosaimh Ó Fátharta (11:44):

Counting people in stations and clusters. You can do anything. We're looking at cameras. So I've been using lenses for years. So some of the engineering companies and serving companies, really good top of the range lenses. And they can monitor movement for me. So picture, you've got like a bridge, and you've got a target on a bridge, and you've got, let's say some sort of a total station measuring the distance. And if any movement, vibration, anything, it'll pick it up. Lo and behold, that lens, the Leica lens, is also able to catch it now without any targets. It'll just catch the corner of a limestone keystone on the bridge, and it will know if it's moved. It'll catch the length of an embankment if it's going straight, and if you have a slide or deviation, it will tell me, oh, it's gone five mls, it's gone 10 mls. Something, what, 10 years ago was...

Richard Donovan (12:36):

Not possible. And it's now possible in very cheap hardware. So the ubiquity of cheap, affordable, high quality hardware is a game changer. The ability to gather the data is the bit, that's probably the bit that's missing that Grafana is filling the gap in. And this is where it starts to become a game changer for say, not so technical, not so software oriented companies to start going, this is doable for us. This is also doable for us to do mostly in-house. And we haven't talked about the stack very much, but obviously, Grafana is a key part of it. We wouldn't have, we don't have a big, I have one graduate, I have no software developers to work with me on this.

Seosaimh Ó Fátharta (13:11):

This is a team of three. That developed this.

Richard Donovan (13:13):

And so started with something like Postgres, and then went up to TimescaleDB because that's a familiar technology. And this is, whilst Grafana has a great stack, we were able, it's so pluggable and adaptable, we're able to plug in initially, the technology that really works with what you've got on the ground. It's not prescribing a very complicated stack. If I had gone in with the full stack of Loki and Mimir and everything else, I would've just lost most of the audience. And I've lost my own graduate working on it. But it can do plain SQL on TimescaleDB. Can plug it into APIs and the simple devices we had, and suddenly, we were up and running with a robust industrial stack, but super simple.

Seosaimh Ó Fátharta (13:55):

Okay, so IRIS itself, so turning data into vision was our kind of catchphrase at the beginning. And I can ally say that we did it. So its foundations are four key pillars of collecting data, monitoring alerts, insights, and decisions. So like I said earlier on, collecting the data was quite easy. We used multiple IoT devices in the field, all remotely connected with SIM cards, and onto the cloud computing. What it allows us to do is to collect real time information, trigger alarms and predefined thresholds that are already set, and send SMS messages to the users, or whoever the maintenance lads are on the ground. Of course, our field, being a big huge organization, customization was a massive part of this, because there's no point sending a text message alert to somebody in one part of the country when it should be going to somebody in the other part.

(14:48):

Also, in Ireland there's 28 days, holidays a year. So it's kind of a managing the guys who are on call, or who have moved from one job to the next was a major thing. Getting their phone numbers, their email addresses in. The alerting system on IRIS allowed us to do that.

(15:06):

The data that we were getting, giving us brilliant insights. Insights we never had before. We can understand the behavior and the relationships between our assets, between our bridges, between our track. Even for the weather, like, when we had flood events. Where am I going with this? So the, yeah, so we had customized dashboards for each area. All that decision allowed us to be really, really reactive when those threshold levels were triggered, and it allowed people to be better decision makers. Like one of the things that everybody kept telling me was, oh, we're all engineers. You know, we want to have engineering judgment. And my answer to them was always, but I'm giving you more time to do that. So let the data guide you and free you up and you won't be trapped looking at Excel charts all day. And that was probably great feedback. So this is what it looks like.

(15:56):

Again,

(15:58):

sorry, it's white. For now. And it'll be black tomorrow. But this is just a dashboard from one of our operators, and his area has, I think it's five different ones there, but straight away, it's all green. And green is gold. Green is good. Green is don't worry about it. But you can see the number four at the bottom, not really jumping out at you, but it's there. And the red warnings, and then there's one amber. And that's what I wanted to hit on. I want 'em to trigger onto that message going, where's that for, and what do I do about it? So let's go to the next slide.

Richard Donovan (16:31):

Yeah. But this, and this, so this is a very basic presentation, very basic stack. We've got, again, one graduate working on this. We haven't really done much with the alerting components. We are using Windows to make sure that the end users are not getting spammed with SMSes, which is also very important. We've got a long roadmap to go, 'cause we're getting now, of course, buy-in and more resources. But we are able to start super simple, and get real value out to the people out and around the world. What's happened to that one?

Seosaimh Ó Fátharta (16:56):

I hit it again.

(16:57):

So here are two different pages, planes that we have. So GIS, that's another in engineering, we love our GIS we love our coordinates, and we love to know where we are. As engineers, we always go to the map. We just go to a map, a chart, whatever you wanna call it. So we've got three red alerts here, and straight away I can tell that there's actually four on the map. So they're scattered around Ireland. Now, there are four different individual areas. So there might have been some work happening that night. The manager comes in the morning, he gets a handover, and he goes, well, why, why have I got red dots on my dashboard? And straight away, he's able to call his local representative in that area and go, we were out there last night and the door wouldn't close on whatever unit it was, and it was giving us an error feed. All those triggers were coming back into Grafana. And we were able to and go and report on it.

(17:51):

Obviously, the reds will have to be fixed straight away, and there's a time limit on them. And the amber is also handy because some things are out of our control. Like it might be, it might be batteries getting low. You know you've got 20, you know, I dunno, 20 hours to get out there to repair the battery before it goes red. And that's really handy. It's almost better the amber stuff than the red stuff because I don't want anything going red. But when it goes amber, I can go and solve the problem before it gets a problem.

Richard Donovan (18:18):

Yep. And that's the snowball that we're getting on. So first challenge we get is the GIS, can we work with GIS and it does, again, we're not buying a super expensive system. Second thing, can we bring in asset data? So everything's driven by numbers and driven by SAP. And what can we pull in asset data and display it so it's relevant to people, and we can, again, for a very low cost, starting with open source system, which is a game changer for them. And as soon as you see this, and as soon as we get the engagement with the users, it's like, okay, we can give 'em the status hot or cold, or what the temperature is. But we can, as you said, we can also start saying, well what's the battery like? When has it being patched? What's the firmware? And it opens up all those conversations in a nice way.

(18:57):

We're not just traditional procurement, we're not generating a massive bunch of requirements to hit people with a stick and giving no solution. With the stack, with Grafana, we're giving them a solution saying, we can gather the data, we can display the data, and now we can start enhancing it and enriching it without starting another massive project that builds another separate system that silos all the data into somewhere else.

Seosaimh Ó Fátharta (19:21):

Yeah. Like, the real time aspect of it is perfect. Like, this one happens in Ireland all the time. Rain. Flooding

(19:31):

And you'd be surprised at how they elevate. So you could like, wake up in the morning, and it's lashing rain. What you don't know it's been lashing the previous two days, and that's building up saturation level, and all of a sudden, if it gets higher than the rail, the top of the rail, I've got to go and put a mitigation in place for that. Before, that involves somebody going out, or be retrospectively informed by the train drivers. Hey, you've got flooding on both sides of the track. Now, we can go and use remote monitoring, put out temporary devices anywhere we want in a matter of hours, and have instant feedback on it coming back to the office. Like, talking about saving time. This is saving travel time for for the operators, getting out in the vans, walking down the tracks, setting up a secure method to get to site, 'cause we're work in red zone. Like, the railways a no-go area.

(20:18):

If you wanna get out there, like, I won't tell you how much paperwork you have to do just to get out there, but it's near to impossible. So having something remote takes the boots off the ground and our safety department will love us first.

Richard Donovan (20:28):

It's 'cause taking possession of a track or going out onto a track can be a dangerous thing. Depending where it is, there could be high speed trains running around. So it's a non-trivial activity to send somebody out to a location to do an inspection. So again, all of this helps that in a very nice, elegant, clean way.

Seosaimh Ó Fátharta (20:44):

On the right hand side of the screen is a screenshot of my phone with the very first alert. Since then, I've discovered that I'm bombing everybody with alerts, and like anybody, once you start sending them too much, they're like, oh, it's another alert from Seosaimh. Another alert, another alert. I quickly found out to grab their attention, and this is to the developers in the room, that I've got three requests from the alert system. The first alert is the initial result. Flashing in red. Then there's repeat alerts every hour or every two hours, or whenever I want to, maybe in orange. And then the best alert is when the problem is solved, and that goes in, I dunno, a line. I won't use green again. But are you with me? It's all, just wanna flash on their phone, have a look at it, and see something quickly.

(21:34):

They don't wanna go looking at the dashboards. The dashboards are great

(21:38):

for my meetings, but the lads who are under pressure tell me the information, I wanna go and react straight away. That's their job. They're like firefighters. And without them, our trains wouldn't move. The level above them, again, in senior management, they love dashboards. They can instantly say, huh? So you gonna work on that? This is this is what happens, you know, after the shift. We gotta stop that. You know, so dashboards are for everybody. Dashboards are for the maintenance guy who's in the van who's got his mobile phone. It's for his manager who can go and push notifications towards him. It's for us in the back office to be able to use the information and push notifications. To create work orders. And then, something happened last week, and I got a phone call from one of our engineers and he asked me, "Seosaimh, could you tell me the temperature of the track on the 15th of August last year?" And I'm like, eh, unusual request, but yeah, give me two seconds. And I went back into the history, and there was another event that day, and he just wanted to know, did the spike in temperature cause it?

(22:37):

Before, I would've had to go back to the vendor, asked him for a check.

Richard Donovan (22:43):

Give him an envelope.

Seosaimh Ó Fátharta (22:44):

Probably would've asked for money. But yeah, I would've asked him for the the trace, because I wouldn't have had it. Having the history on our own server allowed me to go back, and in Grafana, and with the time map, I think I had it in like, I dunno, less than 30 seconds.

Richard Donovan (22:57):

Yeah.

Seosaimh Ó Fátharta (22:58):

And it was great.

Richard Donovan (22:59):

So that's another part of it. Again, we don't sort of pitch it but you start to get sovereignty of your own data, which is very good for these organizations. We're starting to, from the super simple stack, which is a great way to get going, you know, from the sensors, we've moving towards MQTT. So MQTT allows us to start setting a standard across the whole organization, an ontology of all the data, if you will, that actually maps back to SAP, which they absolutely love. That's a big game changer, that we have this correlation of all the data. We're using Timescale for now, because it's, you know, super simple and robust, and is working really well for us. And then we're, you know, diving now deeper into Grafana for the alerting the anomalies. The AI we're very excited about because it provides real operational insight, not just some noise. And the thing we're just tiptoeing into is what we'll talk about is, we don't have time, is sensor fusion.

(23:50):

So when there is rain, and we can see the weather forecast, and we know the status of the tracks, suddenly that becomes a much richer picture for people who are now making decisions about what's really gonna happen. So where do we go? We went from one sensor on a laptop, some demos, then to small operators that would actually talk to us that we were working with that would give us access to some data to basically now built into the procurement contracts that we're going forward with with everybody that they must share the data. We prefer to have an MQTT, but we'll work with almost anything. It's rolled out across regions and hubs across the whole country. It's a simple, robust architecture that scales well with us, especially within MQTT. And we basically have national coverage in a fraction of the time, at fraction of the cost with very happy customers, which is probably-

Seosaimh Ó Fátharta (24:38):

Yeah, really happy.

Richard Donovan (24:38):

Most important thing with it.

(24:42):

We're then working on more integrations now to the service management systems, to the asset management systems. Nearly everything has a service life and a history built into it. We can start working with that data, pulling it from a different direction to give people insights of when they should do their maintenance and their preventative maintenance going on. And for the actual passengers and the customers we're working with, we feel that we're making a real difference in that they should not see the problems. We should know about the problems long before any interruptions of service for customers or threat warnings, and our time to notification, our a time to fix, our a time to be aware of it has gone from potentially literally hours and days. Some random person ringing up, a driver calling in saying, hey, the track is getting flooded, or there's a disruption, to devices going off, real time sensors.

Seosaimh Ó Fátharta (25:33):

Yeah, time. Time is definitely the biggest factor of what the outcome was. Like, I can't emphasize it enough, from getting that phone call to knowing that phone call never has to happen is magic for us. You know? You get on the train in the morning, and you sit down, and you don't know there's a fault, like, I dunno, 50 miles up the road. By the time you get there, I've already known the fault, there's an engineer gone out, solved it, and the train keeps going at line speed. But that train might have been slowed down to travel at a safer, reduced speed. Or it might just be a technical glitch that can be resolved remotely. But if you don't know about it, you can't fix it.

Richard Donovan (26:13):

So that makes us go obviously from very reactive to ours now getting into almost predictive where we can solve things before they actually happen, or nobody's really aware of it. And so for us now, Grafana as a core has become much more than just a dashboard of data that sits on a wall somewhere. It is really an operational hub for information that takes a very old company to a real time data-driven company without having to have massive consultants, and a whole transformation, cultural change. It's that ease of use of work with Grafana. It's a flexibility to deal with feedback that has made it like, kind of so special to us. Not to mention it's open source, which we all love here. It's not taken us down strange roots with strange companies with strange technologies. It's being able to scale with us to now hundreds and thousands of sensors in the roadmap with no problem.

(27:06):

We're running it on Azure for now. It's been super simple to run, it's been robust. We've had zero issues with it, which is, again, been a kind of a revelation for them. Our roadmap, you can kind of guess, it's gonna be AI in there, it's gonna be anomaly detection, it's gonna be insights. We have a backlog of many, many thousands of different types of sensors that people are interested in and fusing them. So what we get an alert, a camera can look at that device, and see is somebody tampering it with a it, has anything changed? Is there a truck hanging off a bridge somewhere 'cause somebody's driven into it? That gets us to a much better place, quicker, faster, better, and easier for the people working with it. They're not out in the field in all weathers trying to find out what the hell happened in a location.

(27:50):

So the kind of takeaways from all of this, with this stack being so flexible, you can start simple. You can start doing showcases, you can start from your laptops. You can be sort of super iterative with your potential customers, clients. and people out there to workshop it in real time. And we drag and drop dashboards in real time with them, which is a game changer for them. It allows you to scale into various technologies that will give you an architecture that will scale out to the real world and take sense of devices from anywhere we've seen,

Seosaimh Ó Fátharta (28:23):

But once we build it, we can just duplicate it. That's the biggest learning I have. Duplication, that loop that goes around, the more sensors I get. Like, I'm telling Richard, we're gonna run out of space, and he's like, ah, don't worry about it. There's plenty of space out there. But yeah, look, it's been great. I can't applaud the guy beside me enough for introducing me to Grafana at the very first day. And we hope to-

Richard Donovan (28:48):

Keep on going, 'because it becomes our sovereign platform for our data, which we can do what we need to and get that insight, and then we're very interested in the Grafana insights as well. How do we start getting more insights out of the data we're collecting? And we go from alerts to much more information across the system.

Seosaimh Ó Fátharta (29:03):

And the business intelligence part of it.

Richard Donovan (29:05):

Yeah. So cool.

Seosaimh Ó Fátharta (29:07):

Okay, thanks very much.