text stringlengths 0 2.35k |
|---|
One of the exercises we did when we have these alert fatigues and like a thousand-odd alerts is we consciously stopped some of the alerts to see who will start shouting. And it happened that more than 50% of these alerts when we turned off, no one actually shouted at us. Like, "Was that even important?" Going through t... |
\[43:16\] I think teams need to be conscious that it's okay to miss a faulty alert, compared to missing out on a real alert which would have cost us millions of pounds, or whatever it is. So I think it's being careful to put the right alerts in, and stopping at that, and not just going overboard with it. Let's take an ... |
**Matt Toback:** I wonder if you could do a -- what would be the equivalent of a bug bounty for alerts? How do incent people to go and clean those up and celebrate being like "They're gone!" |
**Nayana Shetty:** Oh, that's hard. What I've done in this is actually gone and turned them off myself, and being like "Let's see who's going to shout." |
**Matt Toback:** Yeah. \[laughs\] |
**Nayana Shetty:** And when no one shouts, you know that they're not important enough. So that is something that I've done, but I don't know how you would -- hm, interesting... Do you have any ideas? |
**Mat Ryer:** We need some analytics, don't we, on the usage of it, really... Then we all can say "No one's looked at these alerts for ages." Or you could put a specific time on it if you want. I don't wanna design the application now... But yeah, something like that. |
I like the idea that you should go back and look at them and pay attention to whether you still need them, and things. This is a little bit like how in GitHub or in your project management tool - if you have loads of stuff in there, most of it is just getting ignored. And in a way, it creates also this idea that you're... |
So it is that thing of it's just so much there, it stops being useful. I like the idea that -- does it take experience though, do you think, to know what's useful and what's not? |
**Nayana Shetty:** It could do... I mean, the more you see these things, you realize where it's useful and where not... And this is where I've seen some of the junior engineers and teams struggle, which is like - they start worrying about every single alert that comes on Slack, or whichever is your preferred tool, and ... |
So I think it's something that a team should do on a regular basis as like a team activity or something like that, where they sanity-check their alerts. |
One way we used to do it is any alert that we actually did anything with, we started putting some -- all of our alerts used to come to Slack, so we started putting some emojis on it, so we know which of them were actually things that mattered. And on a weekly basis we were like "Oh, there were N of these, which we did ... |
**Matt Toback:** Mat, you need to collect the emojis and then feed it back. Does the IP go in the other way on emojis? |
**Mat Ryer:** We do that in the Grafana Incident tool, but I need to tell the on-call team about that idea. That's such a good idea, because... |
**Matt Toback:** Yeah, that is kind of fun. |
**Mat Ryer:** ...you could collect that data and literally you then -- "Oh, hello!" |
**Matt Toback:** You're like, "What does dancing penguin mean?" \[laughter\] |
**Mat Ryer:** Yeah. It means it's cool. Who owns that idea, legally? I don't know... |
**Nayana Shetty:** I do, because I-- \[laughter\] |
**Mat Ryer:** It's a good idea. |
**Matt Toback:** \[46:59\] Solved! \[laughter\] I wanted to quickly touch on - as you progress in your career, often you're going to walk into new organizations. And you're new to Lego... And everything today is just reinforced this thought that I was kind of \[unintelligible 00:47:13.13\] on this morning, which is all... |
I guess I'm curious - how much about open source tooling makes it easier to transfer into a new organization, and even just how much -- how do you approach going into a new org having this experience, but then also not understanding how everything fits together? |
**Nayana Shetty:** I mean, this is so relatable... I'm going through this now, given I've been in the Lego Group for only three months... And I care about monitoring and in general sustainability of products quite a lot. I've been looking at different teams doing this and thinking "Okay, this team has this Grafana dash... |
I think for me, I was lucky that I was in a team that was building monitoring tools as a service, providing monitoring tools as a service to other teams, so for me it was easier to catch on to what is happening in different areas within the Lego Group. But I think what I fall back to is always think about what are the ... |
On the other side, there's a lot of community that can help you getting started with the tool as well. So I think those are reasons why I would prefer -- like, when suggesting to teams, I would prefer open source technologies when it comes to this space. |
As an example, when I was doing some experimentation for my own personal project, I could have gone with one of the tools that was already available in the organization where I was working, but then I was like "I mean, if I left this organization, I can't take that tool with me, so it's better to have it on more open s... |
**Matt Toback:** Good choice. |
**Nayana Shetty:** But while there is transferable skills within the organization -- like, as an example, the Lego technology, I think we have around 200 or 250-odd teams in it, and if these people within the teams have to move between each other, speaking a common language is quite important... And having that communi... |
I have come from an organization where we were a very big advocate of open source technologies, so I probably would be singing the song of "Let's go all-in on open source." \[laughter\] |
**Mat Ryer:** I'm interested then what's next, what's coming up, and how do you keep your finger on the pulse of what's going on? |
**Nayana Shetty:** I often think less about tools and more about the capabilities that we really need in an organization. It could be anything from like what do we need in terms of system infrastructure side of things, or like the topic for today, more around observability. Around observability I often think the capabi... |
\[51:22\] I have often leaned towards what's happening in the DevOps communities or like in the monitoring communities to actually get insights from them, saying "Oh, there's--" I think two years ago was when I was introduced to Loki, which is the logging tool, and I got super-excited about this, mainly because we were... |
Within the monitoring space, I think my preference within logging would be -- like, if you are in the AWS line, then something like AWS CloudWatch, or Loki, Splunk... These are a couple of tools that I have used in the logging space. You can use the same kind of tools for metrics as well, but there are better tools for... |
And then it's also to do with your metrics aggregation. So you have all of these different metrics and logs and everything that you're collecting, but how do you visualize them together? So you need something around the visualization layer, which is where Grafana or Kibana, one of these kind of tools would come in hand... |
And finally, I think it's the whole -- you're doing all of this because you want alerting. So what do we have in the alerting space? Thinking about "Do I use Slack notifications? Do I use email notifications? SMSes? Pager duty?", whatever. And it's just making sure you understand the capabilities of what you're trying ... |
**Matt Toback:** Yeah. And what I really like about this is -- that's all kind of like through the lens -- or at least the way you think about it as like what are the problems that you're trying to solve for the customer, and what is the value that you're trying to provide. So even at that point, all of these could bec... |
**Nayana Shetty:** Yeah. And also within teams. When there are so many different teams, and every team is autonomous to use their own tools... I think you need the core principles to be the same, irrespective of what tools they're using, and that's where I find having the capabilities and the principle layer set right ... |
**Mat Ryer:** That's amazing. Loads of great practical advice there. And you yourself have spoken at conferences on this subject as well. We'll put one in the show notes for people interested, but you can also google - or use any search engine; other search engines are available. You can duckduckgo and find -- just for... |
Well, unfortunately that's all the time we have, I'm afraid... Thank you so much - Matt Toback was here... Weren't you, Matt? |
**Matt Toback:** I was. Is there anything that you wanna say to your future self, just before we go? |
**Nayana Shetty:** Oh, I love that. I think it is "You don't have to solve all problems today. There are things that you can leave for the future to solve." \[laughs\] That is what I would say to my future self. |
**Mat Ryer:** \[55:33\] Hm. It'd be more useful if we could send messages to our past selves though, because we could tell them what the stocks are gonna do... |
**Matt Toback:** No... Because we know what happens, right? The space-time continuum... It's just not good. |
**Mat Ryer:** Biff gets it doesn't he? Okay. We won't do it then. We won't do it then, Matt. Okay. Well, yes, thank you so much. |
**Nayana Shetty:** I've enjoyed it. This is a topic that I enjoy speaking in general, so... |
**Mat Ryer:** Are you speaking at any other conferences coming up? Or will you? |
**Nayana Shetty:** No. Because I'm new to the organization, I'm just stepping away from speaking, so that I can gather some insights from the organization before I start speaking... So yeah, none. |
**Mat Ryer:** But the videos of your past self are still available, of course, so I do recommend people check them out... |
**Nayana Shetty:** Yeah. |
**Mat Ryer:** Thank you so much to Matt Toback and our special guest, Nayana Shetty. I've been Mat Ryer (still am), and thank you very much for listening. See you next time on Grafana's Big Tent. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.