Connect with attendees on Slack in #grafanacon channel
About this session
Speakers
Dakota (00:00): Hi, we’re here to air out some dirty laundry today and talk about laundry. So who am I? I’m Dakota. I’m still in high school. I think I’m the youngest person in this room. If I’m not, I’d like to find out. I’m from Oregon. I’m currently missing eighth period, and that’s not a real company. I just made that up. And I have a cool graph for you guys, actually. This is why I skip school — to graph things. Boom. Before I get too far into this, is there a point to anything I’m about to tell you? And like most of the lightning talks, the answer is no.
(00:45): Well, kind of, if you want to learn, I think that messing around with Grafana and graphing weird things is the best possible way to do it. So I’m assuming all of you know what you should be graphing, right? Latency, traffic, error saturation — google SRE talks, all sorts of stuff about this. The four golden signals, but should is very different than could, and “could” could be anything right? We can graph anything that has a number that changes over time. I think Abraham Lincoln said that? But anyways, we can graph things like how many unopened emails, how many tabs we have open, what our washing machine’s doing. So what does this really look like in practice? Well, in practice, I got a story for us to get here. So once upon a time, a college dorm, a college friend of mine who does computer science on the side for fun, reached out to me and he was like, “My dorm washing machines are internet connected.”
(01:39): And he was like, “Check out this website.” And I checked out the website and you could click on it and you could see how many minutes are remaining on your dryers. Like, “Go get your laundry please. We don’t want it here.” And I’m like, “You can check out your laundry? That’s awesome. Do you think there’s an API?” And this only goes downhill. Commence shenanigans. Like usual. We found an API. We realized that we could get the state of it minutely. So we were initially going to make some sort of game actually. And like, do we buy or sell stocks based off of what washing machines are doing what? But we realized — I realized that wasn’t a very good idea. And then we’re like, “Oh, what if we just graph it?” And bam, we can graph laundry. Here’s the best time to run. It’s between two and 6:00 AM if you can’t tell.
(02:30): And these are 32 machines — 16 washers, 16 dryers, and they’re all in the resident hall of the freshmen dorms. We can also see people switch between the loads. You can see up, somebody starts to washer right here, then you can see them move it to the dryer. And then when we were doing this and we were exploring their API, which was unsecured unfortunately, we realized there was other sick data fields we could get, whether the was open or closed or the soil level. Unfortunately we couldn’t graph how long stuff’s sat in the machine in the limbo between somebody changing it because I live out of the dryer. I don’t know about you guys.
(03:14): So since October of last year, we have about 9 million rows worth of data of just washers and dryers in our Timescale database. So here’s a bigger graph. Whole dashboard, right? Like wow, lots of things. So we can see how many loads we’ve run. We can see the total usage of if the washer was running nonstop 24/7. We can get current state. There’s all sorts of fun little metrics we have. And here’s more graphs too, right? I love graphs. We can go back and forth between what’s up and what’s down in a sense. And if you want a more technical writeup, this is my buddy’s writeup. He also needs a job just like me.
(03:53): You should totally check this out. His blog is so much cooler than mine or more technical and maybe less funny. Anyways, we’re going to pivot a little and segue over to how many, I can’t see anything, but how many tabs do you think you guys have open? Do you think you have more than a hundred? Open? Raise your hand if you think I have more than a hundred tabs open right now. (Heckler: You can’t see!) I can see a little. Do you think you have more than 500 tabs open right now? A thousand? Well, you see where this is going? Boop, graphed it.
(04:26): So you can see when I program and then I stop programming, right? I close out my tabs because I have problems if I don’t. And we can graph tabs. This was just a Chrome extension actually. And you can also see the outages in my home lab. I don’t have very good power. And when I was in there making my Chrome extension because I was like, “This is fun. What other metrics can I grab?” A lot apparently: Bookmarks, how much RAM is available. I didn’t know they could even get that. Muted tabs, total windows. So you see where this is going, right? BAM! Dashboard.
(05:01): You can download this extension. Actually you can contribute to my dashboard if you’d like to. And you can join the public tab graph or you can run your own private one. I set that all up too and wrote documentation on it. So if you feel like graphing tabs, you totally should. So we’re going to pivot again because why not? And here’s a graph of all the graphs I don’t regret making, right? I think everyone should graph something dumb. We spend all day at work graphing computers and how many seconds it takes to respond to an error. I think the most exciting things you can graph are the things that are real. I spent maybe five minutes the first day, someone can tell you about it, of can we graph how many coffees people are getting? Unfortunately, no. But we were close. Really my process is, I just like to find fun things to graph. I do it outside at the gym anywhere I go, listening to friends’ bad app ideas, lots of graphs come from that.
(05:57): And you might be asking, well, what’s your stack, bro? Well, I’m proud to introduce the G.O.P.T. stack. It makes use of Grafana, Go, Prometheus and timescale. And it looks something like this when you use it. I write a lot of my scraping scripts that scrape from unauthenticated APIs in Go. And depending on my mood and/or energy levels, it either gets pushed into a timescale database for really weird advanced queries and then Prometheus when I’m tired and lazy. But it’s still amazing to work with. And from that, well, we graph it all in Grafana, right? Push gateway is pretty awesome. It was a big player. When you have something, when you want to push from thin clients outwards like a browser extension for how many tabs you have open, I have plenty of time. Let’s see what else we got. So what comes next? I’m hoping we can find patterns in this arbitrary data between things like how much rainfall is there to how much drier usage we have? I think that there’s a connection between all the data that we ever have, right? Points upon points. And also I find an internship for this summer. I’ll be at the experts booth after this if anyone’s interested. I am for hire. I’ll fly anywhere. Anywhere.
(07:15): So from that, I was always told to do a “So what?” by my language arts teacher. So with that being — key takeaways. You should run laundry between two and 6:00 AM If you’re in college, I promise you it’s the best time. Statistically, I have months of data to prove it. And just we’re so preoccupied with whether or not we could do something we sometimes should stop. Whether or not should we. If you know where that quote’s from, you’re good. Umm, man, I’m under like usual. Anyways. That’s all I have today. If you like how I talk and you think it’s funny, I kind of write my blog the same way (codaea.com), or at least I try to the most part. And hopefully you learn something and you all laughed. So I accomplished rule number one. Thank you.