1 00:00:00,000 --> 00:00:03,083 We last spoke in December 2021, which is 2 00:00:03,083 --> 00:00:04,166 almost three years ago. 3 00:00:04,500 --> 00:00:06,583 And the title of that conversation was 4 00:00:06,583 --> 00:00:07,833 Keep on-call simple 5 00:00:08,791 --> 00:00:09,791 And you check it out at 6 00:00:09,791 --> 00:00:11,791 shipit.show/36 7 00:00:12,375 --> 00:00:13,625 So I'm wondering what is new? 8 00:00:14,083 --> 00:00:15,583 Actually, everything new 9 00:00:15,583 --> 00:00:18,125 except the word "Keep". 10 00:00:18,541 --> 00:00:21,500 Because "Keep" is the name of a new 11 00:00:21,500 --> 00:00:23,541 startup we are speaking today about. 12 00:00:24,750 --> 00:00:26,916 And it's no more on call. 13 00:00:27,708 --> 00:00:31,333 It's still trying to make 14 00:00:31,333 --> 00:00:33,041 it simple, keep it simple. 15 00:00:33,958 --> 00:00:37,166 And actually, if like last time we talked 16 00:00:37,166 --> 00:00:40,458 about Grafana Labs, how Grafana Labs 17 00:00:40,458 --> 00:00:43,000 acquired the startup I built together 18 00:00:43,000 --> 00:00:44,125 with my co-founder, 19 00:00:44,125 --> 00:00:46,083 Ildar, back in the days. 20 00:00:46,750 --> 00:00:49,416 And this startup was dedicated to 21 00:00:49,583 --> 00:00:51,875 help on call engineers 22 00:00:51,875 --> 00:00:53,750 set up on call rotations, 23 00:00:53,750 --> 00:00:55,625 set up on call escalations, 24 00:00:56,333 --> 00:01:00,125 And Keep, the new startup 25 00:01:00,125 --> 00:01:01,208 we're talking about today. 26 00:01:02,125 --> 00:01:05,083 It's not focusing on who to notify, 27 00:01:05,083 --> 00:01:06,791 and how to notify, 28 00:01:06,791 --> 00:01:09,416 it's focusing on what. 29 00:01:09,416 --> 00:01:10,916 So we also have Tal joining us today. 30 00:01:11,166 --> 00:01:12,958 I tell Matvey all the time that I think 31 00:01:12,958 --> 00:01:16,666 he's super humble because he managed to 32 00:01:16,666 --> 00:01:20,166 completely change the 33 00:01:20,166 --> 00:01:22,458 whole market of IRM, 34 00:01:22,500 --> 00:01:24,416 Incident Response Management tools. 35 00:01:25,208 --> 00:01:29,125 I think since they built Amixr, which 36 00:01:29,125 --> 00:01:31,666 was acquired by Grafana Labs and today is 37 00:01:31,666 --> 00:01:33,875 known as Grafana OnCall. 38 00:01:34,125 --> 00:01:37,708 There was a huge shift in that market. 39 00:01:38,250 --> 00:01:40,416 We saw tons of new 40 00:01:40,416 --> 00:01:42,541 players coming into that field. 41 00:01:42,541 --> 00:01:46,458 And I also think that it's somewhat the 42 00:01:46,458 --> 00:01:50,750 grounds for what Keep is building to be. 43 00:01:51,208 --> 00:01:53,458 My last job before I became 44 00:01:53,458 --> 00:01:55,166 a startup founder was 45 00:01:55,166 --> 00:01:56,333 an engineer on call. 46 00:01:56,875 --> 00:01:59,125 So I have a lot of empathy to everything 47 00:01:59,125 --> 00:02:01,208 happening during incidents, 48 00:02:01,208 --> 00:02:02,916 before incidents, after incidents, 49 00:02:02,958 --> 00:02:04,583 and especially with people 50 00:02:04,583 --> 00:02:06,375 who are handling incidents. 51 00:02:07,500 --> 00:02:09,125 And when I think about what 52 00:02:09,125 --> 00:02:11,000 what I want to work on 53 00:02:11,000 --> 00:02:12,041 is the first thing 54 00:02:12,041 --> 00:02:13,041 which comes to my mind. 55 00:02:13,583 --> 00:02:17,083 And I'm lucky I met Tal and 56 00:02:17,083 --> 00:02:19,708 our third co-founder, Shahar, who 57 00:02:19,708 --> 00:02:22,375 we share this experience with. 58 00:02:22,875 --> 00:02:25,875 It's safe to say that I guess 59 00:02:25,875 --> 00:02:28,500 every engineer or someone with 60 00:02:28,500 --> 00:02:31,208 engineering background shares the same, 61 00:02:31,333 --> 00:02:34,333 at least perspective about this world. 62 00:02:36,250 --> 00:02:39,291 Before I met Matvey, what brought me 63 00:02:39,291 --> 00:02:42,083 into this world was the exact same thing 64 00:02:42,083 --> 00:02:43,541 as Matvey just mentioned. 65 00:02:44,958 --> 00:02:46,458 Being an engineer, 66 00:02:46,458 --> 00:02:48,375 being an engineering manager, 67 00:02:48,375 --> 00:02:51,583 facing incidents, 68 00:02:51,583 --> 00:02:53,833 facing alert fatigue, 69 00:02:53,833 --> 00:02:56,416 muting different Slack channels, 70 00:02:56,416 --> 00:02:57,500 all of that. 71 00:02:58,000 --> 00:02:59,875 So if you were to take us through an 72 00:02:59,875 --> 00:03:03,000 incident that had all the 73 00:03:03,000 --> 00:03:04,083 things about it were wrong. 74 00:03:04,500 --> 00:03:06,625 How it started, how it happened, how it 75 00:03:06,625 --> 00:03:08,500 unfolded, how it eventually 76 00:03:08,500 --> 00:03:10,291 ended and the follow through. 77 00:03:10,750 --> 00:03:12,958 Do you have such an example in mind, Tal? 78 00:03:14,500 --> 00:03:16,208 I have a few examples just 79 00:03:16,208 --> 00:03:19,291 from the last couple of months. 80 00:03:19,458 --> 00:03:21,833 CrowdStrike is a very large, 81 00:03:22,791 --> 00:03:25,458 I can say antivirus company. 82 00:03:25,958 --> 00:03:30,041 It has agents installed on endpoints in 83 00:03:30,041 --> 00:03:30,916 different companies. 84 00:03:31,000 --> 00:03:33,416 They released some update to their 85 00:03:33,416 --> 00:03:35,083 software that caused 86 00:03:35,083 --> 00:03:38,458 basically all the computers to crash. 87 00:03:39,916 --> 00:03:41,708 Blue screen, if that 88 00:03:41,708 --> 00:03:44,666 says something to somebody. 89 00:03:45,041 --> 00:03:47,000 And I think what's fascinating about this 90 00:03:47,000 --> 00:03:50,333 incident is that it actually caused 91 00:03:50,333 --> 00:03:53,500 hundreds of millions in dollars in loss 92 00:03:53,500 --> 00:03:55,041 for so many companies. 93 00:03:55,583 --> 00:03:57,041 So that's one thing I have 94 00:03:57,041 --> 00:03:59,541 in mind that was just recent. 95 00:03:59,916 --> 00:04:02,375 And the other thing I have in mind is 96 00:04:02,375 --> 00:04:06,250 there was some incident with a major 97 00:04:06,250 --> 00:04:08,958 cloud provider, Google cloud provider, 98 00:04:08,958 --> 00:04:12,416 where they basically made some mistake. 99 00:04:12,416 --> 00:04:16,000 I think it's not even public what was the 100 00:04:16,000 --> 00:04:17,833 main reason for that, but they made some 101 00:04:17,833 --> 00:04:20,041 mistake and they erased a complete 102 00:04:20,041 --> 00:04:21,500 environment for some very 103 00:04:21,500 --> 00:04:23,791 large insurance company in the US, 104 00:04:24,125 --> 00:04:27,166 which also caused like, I think three 105 00:04:27,166 --> 00:04:29,083 days of downtime to that company. 106 00:04:30,041 --> 00:04:32,541 And that was a big thing as well. 107 00:04:32,750 --> 00:04:35,791 And it's the type of things that you as 108 00:04:35,791 --> 00:04:38,333 an engineer, when you think about 109 00:04:38,333 --> 00:04:41,041 the one guideline you have is 110 00:04:41,041 --> 00:04:43,625 in today's cloud environment is that 111 00:04:43,625 --> 00:04:45,916 major cloud providers, they can fail. 112 00:04:46,416 --> 00:04:48,833 And you learn that they actually can. 113 00:04:49,208 --> 00:04:51,958 So I guess those are my two examples. 114 00:04:52,875 --> 00:04:54,375 Everyone makes mistakes, right? 115 00:04:55,041 --> 00:04:57,416 And systems that are able to handle mistakes 116 00:04:57,416 --> 00:04:58,500 they're hard to design 117 00:04:58,500 --> 00:04:59,541 they're hard to maintain 118 00:05:00,291 --> 00:05:02,333 And the more complexity is added 119 00:05:02,333 --> 00:05:04,375 the more difficult this becomes, for sure. 120 00:05:05,125 --> 00:05:07,041 So CrowdStrike, I think that's a great 121 00:05:07,041 --> 00:05:10,166 example because it, first of all, it was 122 00:05:10,166 --> 00:05:12,708 by far, I think, like the biggest outage 123 00:05:12,708 --> 00:05:15,041 in the history of all 124 00:05:15,041 --> 00:05:16,083 information technology. 125 00:05:16,666 --> 00:05:17,416 It was recent. 126 00:05:17,708 --> 00:05:18,125 It was July. 127 00:05:18,666 --> 00:05:20,083 It was just like, as you mentioned, like 128 00:05:20,083 --> 00:05:21,666 some number of months ago, it wasn't even 129 00:05:21,666 --> 00:05:23,250 a year ago that this happened. 130 00:05:23,666 --> 00:05:25,041 It affected everyone. 131 00:05:25,750 --> 00:05:27,166 Government, hospitals, 132 00:05:27,416 --> 00:05:28,833 this was like really serious. 133 00:05:29,625 --> 00:05:31,625 And even though machines crashed that, 134 00:05:31,958 --> 00:05:33,125 you know, like they shouldn't have 135 00:05:33,125 --> 00:05:34,541 crashed, they couldn't restart. 136 00:05:35,375 --> 00:05:36,541 I think that was the big one. 137 00:05:36,916 --> 00:05:37,750 And a fix was rolled 138 00:05:37,750 --> 00:05:38,875 out, you know, within hours. 139 00:05:39,250 --> 00:05:41,250 Like the fix was out, but the fix 140 00:05:41,250 --> 00:05:42,666 couldn't be applied fast enough. 141 00:05:42,666 --> 00:05:44,458 Everyone was in damage control. 142 00:05:45,125 --> 00:05:46,375 It was messy, really messy. 143 00:05:46,916 --> 00:05:47,875 I was looking on Wikipedia. 144 00:05:49,041 --> 00:05:51,625 I was looking this up because I knew that 145 00:05:51,625 --> 00:05:52,750 we would talk about this. 146 00:05:53,375 --> 00:05:56,750 And apparently, the financial damage was 147 00:05:56,750 --> 00:05:59,791 closer to 10 billion dollars worldwide. 148 00:06:00,458 --> 00:06:01,833 And it's an estimate by the financial 149 00:06:01,833 --> 00:06:03,625 institutions, but that's huge. 150 00:06:04,375 --> 00:06:05,250 This reminds me of something that 151 00:06:05,250 --> 00:06:07,541 Matvey mentioned when he first spoke. 152 00:06:07,791 --> 00:06:09,416 He mentioned about the black swans 153 00:06:09,416 --> 00:06:10,541 and the black swan events. 154 00:06:11,250 --> 00:06:11,958 And I think this 155 00:06:11,958 --> 00:06:13,875 definitely classes as one. 156 00:06:14,541 --> 00:06:16,208 So what are your thoughts there, Matvey? 157 00:06:16,708 --> 00:06:19,541 By the way, I just want to throw a fun 158 00:06:19,541 --> 00:06:21,291 fact, but I think it's definitely 159 00:06:21,291 --> 00:06:23,458 a black swan for CrowdStrike. 160 00:06:23,583 --> 00:06:25,916 Because I'm looking at their stock price 161 00:06:25,916 --> 00:06:28,708 from like six months ago. 162 00:06:28,708 --> 00:06:29,666 And it used to be like 163 00:06:29,666 --> 00:06:33,166 almost 400 US dollars per stock. 164 00:06:33,166 --> 00:06:36,375 And it's like less than 300 right now. 165 00:06:36,375 --> 00:06:39,416 So I think for them, it was a big thing. 166 00:06:39,708 --> 00:06:42,416 Other fun fact about CrowdStrike and it's 167 00:06:42,416 --> 00:06:43,916 a little bit closer to us. 168 00:06:45,000 --> 00:06:46,833 It's like a small story. 169 00:06:49,416 --> 00:06:51,791 Since our startup is relatively young 170 00:06:52,500 --> 00:06:54,333 we are selling to 171 00:06:54,333 --> 00:06:56,125 enterprises, large enterprises. 172 00:06:56,958 --> 00:06:57,958 Practically all of them. 173 00:06:59,000 --> 00:07:00,375 I think all of them were affected. 174 00:07:01,833 --> 00:07:05,000 And one of our customersZZZ, 175 00:07:05,000 --> 00:07:07,416 prospects at the time, like customers now, 176 00:07:07,916 --> 00:07:10,458 they wanted to onboard to Keep actually 177 00:07:10,458 --> 00:07:13,083 to mitigate such things faster, 178 00:07:13,458 --> 00:07:15,500 such events faster, such black swans. 179 00:07:16,208 --> 00:07:16,916 Going back to your 180 00:07:16,916 --> 00:07:18,125 question about black swans. 181 00:07:18,958 --> 00:07:20,833 most black swan for me was 182 00:07:20,833 --> 00:07:21,791 major outage 183 00:07:21,791 --> 00:07:23,750 of one of cloud providers 184 00:07:23,750 --> 00:07:25,166 we used at Grafana. 185 00:07:25,750 --> 00:07:29,958 And I can't disclose a lot about it, but it 186 00:07:29,958 --> 00:07:31,833 was a huge surprise for me that, 187 00:07:32,416 --> 00:07:33,791 you know, when you look at 188 00:07:33,791 --> 00:07:35,958 cloud providers' uptime pages 189 00:07:35,958 --> 00:07:37,333 like status pages 190 00:07:37,583 --> 00:07:39,625 you see a lot of items like 191 00:07:40,416 --> 00:07:42,541 SQL service 192 00:07:42,833 --> 00:07:45,250 like hosted SQL, something else, DNS. 193 00:07:45,791 --> 00:07:48,583 And you assume that if they will go down, 194 00:07:49,083 --> 00:07:50,791 like one of them will go down or maybe 195 00:07:50,791 --> 00:07:52,291 two of them will go down. 196 00:07:53,166 --> 00:07:55,833 It's hard for you to imagine that 197 00:07:55,833 --> 00:07:57,958 actually the whole screen could be red. 198 00:07:58,375 --> 00:08:01,666 I think that's the market we operate in. 199 00:08:01,708 --> 00:08:05,416 Like there is a Murphy's Law that says: 200 00:08:07,166 --> 00:08:10,916 everything that can go wrong, will go wrong. 201 00:08:11,333 --> 00:08:15,708 You know, you try to build walls behind 202 00:08:15,708 --> 00:08:19,291 large scale systems, but eventually, 203 00:08:19,291 --> 00:08:21,333 you know, people make mistakes. 204 00:08:21,958 --> 00:08:23,750 Software is not complete 205 00:08:24,125 --> 00:08:26,291 There will always be something happening. 206 00:08:26,291 --> 00:08:29,625 It can be like from electricity going down 207 00:08:29,625 --> 00:08:33,625 in some, I don't know, a server factory 208 00:08:34,416 --> 00:08:36,125 to a line of code that somebody 209 00:08:36,125 --> 00:08:37,666 changes and breaks everything. 210 00:08:38,666 --> 00:08:40,583 Things go wrong 211 00:08:41,958 --> 00:08:43,291 That's one true phrase 212 00:08:43,291 --> 00:08:45,000 you can always rely on. 213 00:08:45,416 --> 00:08:47,208 Yeah, things are not going to work. 214 00:08:47,208 --> 00:08:48,333 What are you going to about it? 215 00:08:48,416 --> 00:08:49,333 I like that take. 216 00:08:49,750 --> 00:08:53,458 So if we think about CrowdStrike and if 217 00:08:53,458 --> 00:08:55,291 we think about having Keep. 218 00:08:55,625 --> 00:08:57,708 So let's assume that the customer 219 00:08:57,708 --> 00:08:58,750 that you mentioned 220 00:08:58,750 --> 00:09:00,125 that is now customer 221 00:09:00,125 --> 00:09:02,291 if they had Keep at the time 222 00:09:02,291 --> 00:09:04,375 before CrowdStrike happened 223 00:09:04,375 --> 00:09:06,125 how would that have changed 224 00:09:06,166 --> 00:09:08,875 how things unfolded for them? 225 00:09:09,458 --> 00:09:13,083 Customers of that scale and I'm speaking 226 00:09:13,083 --> 00:09:16,291 about companies with probably hundreds 227 00:09:16,291 --> 00:09:20,791 and thousands of physical sites worldwide 228 00:09:20,791 --> 00:09:24,125 and dozens of thousands of people working 229 00:09:24,125 --> 00:09:25,625 in those enterprises. 230 00:09:26,208 --> 00:09:31,375 They have multiple solutions to keep 231 00:09:31,375 --> 00:09:32,958 track of what's going on in their 232 00:09:32,958 --> 00:09:35,000 infrastructure, monitoring systems. 233 00:09:35,750 --> 00:09:38,250 Actually, the phrase which we heard on 234 00:09:38,250 --> 00:09:41,000 one of customers calls is that we usually 235 00:09:41,000 --> 00:09:42,833 ask like which monitoring system you use 236 00:09:42,833 --> 00:09:44,875 and they answered as like, actually, 237 00:09:45,458 --> 00:09:47,250 anything you could you could think of. 238 00:09:48,166 --> 00:09:49,375 We have everything in 239 00:09:49,375 --> 00:09:51,458 somewhere in our org. 240 00:09:52,333 --> 00:09:54,125 I think you have some counter for the 241 00:09:54,125 --> 00:09:55,416 number of monitoring 242 00:09:55,416 --> 00:09:56,500 tools out there, right? 243 00:09:56,791 --> 00:10:00,125 I tried to count, like I found 370 plus, 244 00:10:01,250 --> 00:10:04,208 and I kept finding new ones. 245 00:10:04,791 --> 00:10:08,583 So those large enterprises, 246 00:10:08,583 --> 00:10:12,291 they monitor a lot of things, 247 00:10:12,291 --> 00:10:14,791 and all those monitoring systems, they 248 00:10:14,791 --> 00:10:19,416 generate signals, alerts, alarms, 249 00:10:20,291 --> 00:10:21,875 how some people call them. 250 00:10:22,958 --> 00:10:26,000 And the problem is that if they collect 251 00:10:26,000 --> 00:10:27,791 all those signals in one place, 252 00:10:28,333 --> 00:10:30,791 they can't actually manage this scale. 253 00:10:31,375 --> 00:10:36,958 So we talk with people who receive like 254 00:10:36,958 --> 00:10:39,000 thousands and thousands an hour. 255 00:10:40,041 --> 00:10:42,750 There is one particular customer who is 256 00:10:42,750 --> 00:10:45,416 dealing with 70,000 alerts a day. 257 00:10:46,083 --> 00:10:47,875 And the problem they are facing is a 258 00:10:47,875 --> 00:10:50,208 problem of finding the 259 00:10:50,208 --> 00:10:51,291 needle in the haystack. 260 00:10:52,291 --> 00:10:54,083 Like this is an actual incident. 261 00:10:54,916 --> 00:10:55,916 This is not noise. 262 00:10:56,458 --> 00:10:57,625 And the rest is noise. 263 00:10:58,708 --> 00:10:59,541 This is a problem. 264 00:11:00,000 --> 00:11:03,625 We at Keep, we don't solve the problem of 265 00:11:03,625 --> 00:11:05,166 who to notify, and how to notify, 266 00:11:05,416 --> 00:11:06,333 and what to do 267 00:11:06,333 --> 00:11:07,791 actually during the incident. 268 00:11:08,750 --> 00:11:10,000 This is IRM space. 269 00:11:11,500 --> 00:11:14,083 This is the space for such products as 270 00:11:14,083 --> 00:11:15,541 like incident.io 271 00:11:15,875 --> 00:11:18,125 Grafana on call, PagerDuty. 272 00:11:19,250 --> 00:11:22,291 But we're building the best software 273 00:11:22,291 --> 00:11:24,291 which will help you to look at it, 274 00:11:24,708 --> 00:11:26,250 look at this stream of events and say 275 00:11:26,250 --> 00:11:28,041 like, those five events, 276 00:11:28,541 --> 00:11:29,958 they happened in the past. 277 00:11:31,666 --> 00:11:32,791 And most probably, 278 00:11:32,791 --> 00:11:33,750 this is a real incident. 279 00:11:34,041 --> 00:11:36,958 I think people sometimes underestimate 280 00:11:36,958 --> 00:11:38,750 like, you know, with 281 00:11:38,750 --> 00:11:39,875 the CrowdStrike event, 282 00:11:40,125 --> 00:11:46,333 what you as end user get is the report 283 00:11:46,333 --> 00:11:49,291 that CrowdStrike publishes about the 284 00:11:49,291 --> 00:11:50,791 event that happened. 285 00:11:51,583 --> 00:11:55,291 But people underestimate the chaos that 286 00:11:55,291 --> 00:11:56,750 companies went through 287 00:11:57,125 --> 00:12:00,083 to try to understand what happens before 288 00:12:00,083 --> 00:12:01,750 CrowdStrike even realized 289 00:12:01,833 --> 00:12:02,791 that they released 290 00:12:02,791 --> 00:12:04,250 something that breaks everything. 291 00:12:05,041 --> 00:12:07,583 So just think about it as a point in time 292 00:12:07,583 --> 00:12:10,666 where all the screens start to become red 293 00:12:10,958 --> 00:12:14,000 and there are people sitting in, usually 294 00:12:14,000 --> 00:12:15,166 it's like network 295 00:12:15,166 --> 00:12:18,083 operation centers or, you know, 296 00:12:18,250 --> 00:12:20,791 individuals monitoring their systems and 297 00:12:20,791 --> 00:12:22,125 everything becomes red. 298 00:12:22,333 --> 00:12:23,833 And that's a point in time 299 00:12:23,833 --> 00:12:25,708 where you are just in chaos. 300 00:12:26,083 --> 00:12:28,208 And now you need to start figuring out... 301 00:12:28,416 --> 00:12:30,583 Okay, what's happening? 302 00:12:30,916 --> 00:12:32,000 Where does it start? 303 00:12:32,458 --> 00:12:36,250 What's the single point in time where 304 00:12:36,250 --> 00:12:37,750 everything started to collapse? 305 00:12:38,000 --> 00:12:39,458 Why do you think that companies end up in 306 00:12:39,458 --> 00:12:40,666 a situation where they 307 00:12:40,666 --> 00:12:43,541 have a lot of alerts? 308 00:12:44,166 --> 00:12:46,166 Like, how do you end up in that place? 309 00:12:46,708 --> 00:12:48,291 First of all, how do you end up with 310 00:12:48,291 --> 00:12:50,041 having so many monitoring solutions? 311 00:12:50,375 --> 00:12:52,333 I understand two or three, right? 312 00:12:52,333 --> 00:12:54,125 Because you want the monitoring to check 313 00:12:54,125 --> 00:12:55,000 your other monitoring. 314 00:12:55,041 --> 00:12:56,583 So you would have a 315 00:12:56,583 --> 00:12:58,708 few, but more than that? 316 00:13:00,291 --> 00:13:02,250 I have a lot of answers for that. 317 00:13:02,250 --> 00:13:03,791 But one of the things I want to check is 318 00:13:03,791 --> 00:13:06,208 like the number of, I think 319 00:13:06,208 --> 00:13:08,250 Netflix is famous for that, 320 00:13:08,250 --> 00:13:11,041 but the number of microservices they 321 00:13:11,041 --> 00:13:13,416 have, I think it's more 322 00:13:13,416 --> 00:13:15,916 than a thousand microservices 323 00:13:16,375 --> 00:13:17,208 that they have. 324 00:13:17,208 --> 00:13:20,250 They have this famous graph of Netflix 325 00:13:20,250 --> 00:13:21,958 microservices and like 326 00:13:21,958 --> 00:13:24,250 the connections between them 327 00:13:24,541 --> 00:13:26,458 and the different frameworks they're 328 00:13:26,458 --> 00:13:28,250 running, the different 329 00:13:28,250 --> 00:13:30,000 database systems they're using, 330 00:13:30,000 --> 00:13:30,708 the different queue 331 00:13:30,708 --> 00:13:32,000 systems they're using. 332 00:13:32,458 --> 00:13:33,541 It's a modern company, 333 00:13:33,541 --> 00:13:35,000 but it started a while ago. 334 00:13:35,708 --> 00:13:39,750 So just think about, at least that's a 335 00:13:39,750 --> 00:13:41,750 good example from my perspective. 336 00:13:42,250 --> 00:13:44,291 Just think about a company that starts, 337 00:13:44,875 --> 00:13:47,125 it has some infrastructure, 338 00:13:47,125 --> 00:13:49,416 some complexity of stuff. 339 00:13:50,083 --> 00:13:51,250 People join the company. 340 00:13:51,500 --> 00:13:53,500 It's growing. 341 00:13:54,166 --> 00:13:58,125 It has a major growth such as Netflix. 342 00:13:59,250 --> 00:14:00,916 More complex infrastructure. 343 00:14:02,166 --> 00:14:03,875 Team members who joined from different 344 00:14:03,875 --> 00:14:05,708 companies, they like to work with the 345 00:14:05,708 --> 00:14:06,958 tools that they're used to. 346 00:14:07,458 --> 00:14:08,875 New technology comes in. 347 00:14:09,166 --> 00:14:12,208 It has its own monitoring tool, or it 348 00:14:12,208 --> 00:14:14,166 adopts 10 different monitoring tools. 349 00:14:14,916 --> 00:14:17,250 A new guy comes in, he brings his own 350 00:14:17,250 --> 00:14:19,666 methodology of how you do things. 351 00:14:20,458 --> 00:14:23,000 And as the company grows, 352 00:14:23,125 --> 00:14:24,416 a few years later, you find 353 00:14:24,416 --> 00:14:26,250 yourself with people's legacy, 354 00:14:26,583 --> 00:14:28,625 like the things they brought in, the 355 00:14:28,625 --> 00:14:30,791 software they wrote, the 356 00:14:30,791 --> 00:14:32,333 tools that they were using, 357 00:14:33,041 --> 00:14:34,166 eventually that all 358 00:14:34,166 --> 00:14:35,208 sticks with the company. 359 00:14:35,791 --> 00:14:38,333 So you get to a point in larger 360 00:14:38,333 --> 00:14:40,000 enterprises where you just 361 00:14:40,000 --> 00:14:42,041 see everything of everything. 362 00:14:43,166 --> 00:14:43,916 Technical debt is 363 00:14:43,916 --> 00:14:45,666 something that you always aggregate, 364 00:14:46,458 --> 00:14:50,083 and you try to find a balance to, 365 00:14:50,083 --> 00:14:52,625 you have your company, it has 366 00:14:52,625 --> 00:14:55,708 its goals, it's a business. 367 00:14:56,125 --> 00:14:57,083 Eventually you need to 368 00:14:57,083 --> 00:14:58,750 make the business grow. 369 00:14:59,583 --> 00:15:04,375 It's very hard to wrap up or accumulate 370 00:15:04,375 --> 00:15:06,416 the return on 371 00:15:06,416 --> 00:15:08,166 investment you have when you 372 00:15:08,625 --> 00:15:10,416 need to fix those kinds of stuff. 373 00:15:10,708 --> 00:15:13,458 The things that doesn't necessarily push 374 00:15:13,458 --> 00:15:14,541 the business forward, 375 00:15:16,541 --> 00:15:20,166 but actually they have hidden costs. 376 00:15:21,375 --> 00:15:23,958 So I think that's how you find, at least 377 00:15:23,958 --> 00:15:26,125 from my perspective, companies who use 378 00:15:26,625 --> 00:15:28,250 tens of different tools. 379 00:15:29,166 --> 00:15:29,333 Okay. 380 00:15:29,875 --> 00:15:31,916 And does Keep, in this context, mean 381 00:15:31,916 --> 00:15:33,833 keep everything you have, 382 00:15:33,833 --> 00:15:35,458 we'll help you make sense of it? 383 00:15:35,958 --> 00:15:36,833 You mean the name? 384 00:15:37,083 --> 00:15:37,458 Yes. 385 00:15:38,250 --> 00:15:42,500 No, the name actually has a cuter story. 386 00:15:43,458 --> 00:15:46,750 Actually, Shahar, our co-founder, that is 387 00:15:46,750 --> 00:15:51,625 not here, we met each other 13 years ago. 388 00:15:51,958 --> 00:15:53,041 We served together in the 389 00:15:53,041 --> 00:15:54,333 Israeli Defense Forces 390 00:15:54,333 --> 00:15:55,958 and both of us, 391 00:15:55,958 --> 00:15:57,958 we used to play Age of Empires, 392 00:15:58,625 --> 00:16:01,625 and we just took something like a keep is 393 00:16:01,625 --> 00:16:02,458 something that you 394 00:16:02,458 --> 00:16:03,791 have in Age of Empires, 395 00:16:04,041 --> 00:16:06,791 a building that you can build there, and 396 00:16:06,791 --> 00:16:09,333 it was just a random story 397 00:16:09,333 --> 00:16:10,791 that we chose something from 398 00:16:10,791 --> 00:16:11,333 that game. 399 00:16:12,083 --> 00:16:13,041 Oh wow, we can see it. 400 00:16:13,291 --> 00:16:13,875 So that's a keep. 401 00:16:14,375 --> 00:16:15,458 It's like a tower. 402 00:16:16,208 --> 00:16:16,416 Okay. 403 00:16:16,875 --> 00:16:17,916 Is that for archers? 404 00:16:17,916 --> 00:16:19,291 Is that where archers would be? 405 00:16:20,250 --> 00:16:21,125 I think so. 406 00:16:21,375 --> 00:16:22,375 Yeah, I think so. 407 00:16:22,416 --> 00:16:22,666 Okay. 408 00:16:23,083 --> 00:16:24,250 I remember playing the game too. 409 00:16:24,250 --> 00:16:26,125 That was a while back, but you're right, 410 00:16:26,125 --> 00:16:26,958 that was a fun game. 411 00:16:27,708 --> 00:16:29,041 Can you still play today, by the way? 412 00:16:29,291 --> 00:16:30,375 I think you can already 413 00:16:30,375 --> 00:16:31,875 play that in the browser. 414 00:16:31,958 --> 00:16:34,125 So some people migrated the whole game 415 00:16:34,125 --> 00:16:36,125 into the browser, which is kind of cool, 416 00:16:36,125 --> 00:16:36,583 because you don't 417 00:16:36,583 --> 00:16:37,583 have to install anything. 418 00:16:37,958 --> 00:16:40,458 You remember there used to be CDs, and 419 00:16:40,458 --> 00:16:41,666 you needed to have 420 00:16:41,666 --> 00:16:43,500 license keys and everything 421 00:16:43,500 --> 00:16:44,083 like that. 422 00:16:44,083 --> 00:16:46,166 Now you can play Age of Empires in the 423 00:16:46,166 --> 00:16:48,416 web browser, which is quite cool, but 424 00:16:48,416 --> 00:16:49,375 yeah, you can play it. 425 00:16:49,625 --> 00:16:52,416 What is the significance to the like day to day? 426 00:16:52,625 --> 00:16:53,416 So you have this keep. 427 00:16:54,125 --> 00:16:54,916 What is it protecting? 428 00:16:55,125 --> 00:16:56,291 Or what is it watching over? 429 00:16:56,291 --> 00:16:57,083 It's like a watchtower. 430 00:16:57,833 --> 00:16:59,166 Is that what it's supposed to be? 431 00:16:59,791 --> 00:17:01,458 It's protecting people 432 00:17:01,458 --> 00:17:02,958 who deal with alerts. 433 00:17:03,458 --> 00:17:06,208 One of the examples, we had a call with a 434 00:17:06,208 --> 00:17:08,375 small team of four who 435 00:17:08,375 --> 00:17:10,333 managed alerts in their like 436 00:17:10,416 --> 00:17:11,666 internally built system. 437 00:17:12,291 --> 00:17:14,750 And those people practically looking at 438 00:17:14,750 --> 00:17:17,041 alerts multiple hours a day, 439 00:17:17,791 --> 00:17:19,833 staring at those screens and trying to 440 00:17:19,833 --> 00:17:22,250 see like in a matrix movie, like, is 441 00:17:22,250 --> 00:17:23,125 something happening? 442 00:17:23,750 --> 00:17:25,750 This is not a fun job, I guess. 443 00:17:26,208 --> 00:17:29,208 We want those people to be busy with 444 00:17:29,208 --> 00:17:31,458 something maybe more interesting 445 00:17:32,166 --> 00:17:35,708 to apply some best practices to fix some 446 00:17:35,708 --> 00:17:37,875 core reasons or build 447 00:17:37,875 --> 00:17:39,791 something for future. 448 00:17:39,958 --> 00:17:43,958 So I think that people can try and 449 00:17:43,958 --> 00:17:45,500 imagine what this looks 450 00:17:45,500 --> 00:17:47,375 like when it's all set up 451 00:17:47,416 --> 00:17:48,333 and how it works. 452 00:17:48,708 --> 00:17:50,708 Before we go into the demo part, which 453 00:17:50,708 --> 00:17:51,916 I'm very keen to see how 454 00:17:51,916 --> 00:17:52,875 this works out in practice. 455 00:17:53,583 --> 00:17:55,666 But I'm wondering, why would someone pick 456 00:17:55,666 --> 00:17:57,000 this, which for example, 457 00:17:57,000 --> 00:17:59,083 maybe they don't have that 458 00:17:59,083 --> 00:17:59,791 much volume? 459 00:18:00,750 --> 00:18:01,291 What does that mean? 460 00:18:01,291 --> 00:18:02,958 Even if you have maybe tens or hundreds, 461 00:18:03,375 --> 00:18:03,958 you don't need to have 462 00:18:03,958 --> 00:18:05,500 thousands for this to be useful. 463 00:18:06,041 --> 00:18:06,833 Why would someone pick this? 464 00:18:07,666 --> 00:18:12,666 When we started the way of Keep, what we 465 00:18:12,666 --> 00:18:14,541 used to do to get some 466 00:18:14,541 --> 00:18:16,875 feedback from users was to post 467 00:18:16,875 --> 00:18:18,500 stuff on Hacker News. 468 00:18:20,083 --> 00:18:22,583 You usually get the truth 469 00:18:22,583 --> 00:18:24,250 right to your face over there. 470 00:18:25,416 --> 00:18:27,958 And one of the posts that we wrote that 471 00:18:27,958 --> 00:18:29,833 worked very well for us 472 00:18:29,833 --> 00:18:31,791 actually, and brought in very 473 00:18:31,916 --> 00:18:36,708 meaningful feedback was we posted 474 00:18:36,708 --> 00:18:40,083 something that says GitHub actions for 475 00:18:40,083 --> 00:18:41,166 your monitoring tools. 476 00:18:41,625 --> 00:18:44,416 So one of the main capabilities that you 477 00:18:44,416 --> 00:18:47,125 have within Keep is to 478 00:18:47,125 --> 00:18:49,208 create workflows, automations 479 00:18:49,916 --> 00:18:52,208 that are based, like the triggers are 480 00:18:52,208 --> 00:18:54,583 based on alerts that you get on the 481 00:18:54,583 --> 00:18:55,541 alarms, on the events 482 00:18:55,541 --> 00:18:56,583 that go inside Keep. 483 00:18:56,875 --> 00:18:59,708 I think that was super 484 00:18:59,708 --> 00:19:01,875 interesting for a lot of companies. 485 00:19:02,083 --> 00:19:06,458 We actually saw a lot of other monitoring 486 00:19:06,458 --> 00:19:07,833 tools and even 487 00:19:07,833 --> 00:19:10,583 companies from the IRM space 488 00:19:10,583 --> 00:19:13,750 actually implementing 489 00:19:13,750 --> 00:19:17,958 workflows pretty much the same day. 490 00:19:18,291 --> 00:19:21,041 I'm not saying anybody copied or anything 491 00:19:21,041 --> 00:19:22,291 like that, but we 492 00:19:22,291 --> 00:19:23,583 actually saw them posting 493 00:19:24,541 --> 00:19:26,583 like a few days later that they also 494 00:19:26,583 --> 00:19:28,958 support this type of thing. 495 00:19:28,958 --> 00:19:32,083 And I find a lot of the smaller startups, 496 00:19:32,875 --> 00:19:35,500 smaller companies find a 497 00:19:35,500 --> 00:19:39,083 lot of value with just writing 498 00:19:39,166 --> 00:19:39,916 automation. 499 00:19:39,916 --> 00:19:42,125 So before they become this huge 500 00:19:42,125 --> 00:19:47,125 monitoring beast with this sea of alerts 501 00:19:47,125 --> 00:19:48,291 and sea of signals that 502 00:19:48,291 --> 00:19:50,750 they get, they can actually leverage the 503 00:19:50,750 --> 00:19:52,625 automation capabilities 504 00:19:52,625 --> 00:19:54,083 that we have within Keep. 505 00:19:54,375 --> 00:19:57,583 So they can do from the smallest thing of 506 00:19:57,583 --> 00:19:59,916 creating a ticket 507 00:19:59,916 --> 00:20:01,375 when they get some alert 508 00:20:01,833 --> 00:20:02,958 so they can handle it 509 00:20:02,958 --> 00:20:04,083 and fix it for later. 510 00:20:04,750 --> 00:20:07,791 Or if they want to try to see if they can 511 00:20:07,791 --> 00:20:09,958 automatically fix the 512 00:20:09,958 --> 00:20:11,333 issue, maybe it's just 513 00:20:11,333 --> 00:20:12,875 a matter of restarting 514 00:20:12,875 --> 00:20:14,666 something in their Kubernetes cluster. 515 00:20:15,750 --> 00:20:18,833 That's everything you can do with Keep's 516 00:20:18,833 --> 00:20:19,625 workflow automation. 517 00:20:19,833 --> 00:20:21,375 And I think that's what a lot of the 518 00:20:21,375 --> 00:20:22,875 smaller companies find 519 00:20:22,875 --> 00:20:25,583 interesting in what we're building 520 00:20:25,916 --> 00:20:26,583 within Keep. 521 00:20:28,083 --> 00:20:28,708 Okay, okay. 522 00:20:29,000 --> 00:20:29,833 Now that sounds interesting. 523 00:20:30,333 --> 00:20:31,375 Yeah, happy to share the 524 00:20:31,375 --> 00:20:32,708 Hacker News post as well. 525 00:20:33,250 --> 00:20:34,583 Yes, we'll put it in the show notes. 526 00:20:34,875 --> 00:20:35,791 That's a good idea. 527 00:20:35,791 --> 00:20:36,458 That's a good idea. 528 00:20:37,208 --> 00:20:39,583 But again, there's nothing better, I 529 00:20:39,583 --> 00:20:41,083 think, than seeing the thing in action 530 00:20:42,083 --> 00:20:46,708 and seeing its limits as well, where it 531 00:20:46,708 --> 00:20:48,041 maybe stops its edges. 532 00:20:49,041 --> 00:20:53,375 And also, yeah, those AHA moments 533 00:20:53,375 --> 00:20:54,750 that you can only have 534 00:20:54,750 --> 00:20:55,583 when you see something. 535 00:20:55,958 --> 00:20:57,083 So I'm looking forward to that. 536 00:20:57,958 --> 00:21:00,375 It's time for the demo. 537 00:21:00,375 --> 00:21:02,875 I hope it's a live demo, 538 00:21:02,875 --> 00:21:05,750 so I'm a little bit nervous. 539 00:21:06,958 --> 00:21:09,000 Things that can go wrong will go wrong. 540 00:21:09,416 --> 00:21:09,625 Exactly. 541 00:21:09,833 --> 00:21:11,250 Tal is here to support you. 542 00:21:11,500 --> 00:21:12,750 Everything will go wrong Matvey, 543 00:21:12,750 --> 00:21:13,708 so don't worry. 544 00:21:15,291 --> 00:21:18,166 Being honest, I prepared a little bit. 545 00:21:19,166 --> 00:21:21,875 If it will go wrong, it 546 00:21:21,875 --> 00:21:23,625 will be a surprise for me. 547 00:21:23,750 --> 00:21:25,000 Well, we can help you fix it. 548 00:21:25,291 --> 00:21:25,833 How about that? 549 00:21:27,208 --> 00:21:27,416 Nice. 550 00:21:27,500 --> 00:21:27,916 We can try. 551 00:21:27,916 --> 00:21:29,500 Well, we can at least try helping. 552 00:21:30,041 --> 00:21:30,708 I don't know if we will 553 00:21:30,708 --> 00:21:32,583 succeed, but we're here for you. 554 00:21:33,208 --> 00:21:35,583 I want to wrap a little bit 555 00:21:35,583 --> 00:21:37,583 before I jump into the demo 556 00:21:37,666 --> 00:21:43,333 because before that moment, we spoke 557 00:21:43,333 --> 00:21:46,833 about possibilities of 558 00:21:46,833 --> 00:21:50,958 how Keep is able to find 559 00:21:51,041 --> 00:21:53,000 incidents in the lake of alerts. 560 00:21:54,458 --> 00:21:55,666 And this feature, 561 00:21:56,125 --> 00:21:58,083 those AI features we have, 562 00:21:58,583 --> 00:22:00,000 those are actually pretty advanced 563 00:22:00,000 --> 00:22:02,416 features we sell to large enterprises. 564 00:22:03,166 --> 00:22:07,583 And those features, they are based in our 565 00:22:07,583 --> 00:22:09,583 open source Keep, which 566 00:22:09,583 --> 00:22:12,333 is pretty well adopted by 567 00:22:12,333 --> 00:22:15,000 large enterprises and small companies who 568 00:22:15,000 --> 00:22:17,041 don't have this amount of alerts. 569 00:22:17,958 --> 00:22:20,000 And I will speak about 570 00:22:20,000 --> 00:22:21,375 this open source part. 571 00:22:21,875 --> 00:22:26,958 So if you are interested in AI, reach out 572 00:22:26,958 --> 00:22:30,583 to us, but we will not spend your time on 573 00:22:31,583 --> 00:22:32,666 advertisement here. 574 00:22:34,666 --> 00:22:37,166 And this part, this open source part, 575 00:22:37,166 --> 00:22:38,833 it's like a Swiss knife. 576 00:22:39,375 --> 00:22:43,541 So Tal mentioned, it has workflows like 577 00:22:43,541 --> 00:22:44,416 GitHub Actions for 578 00:22:44,416 --> 00:22:45,958 alerts, it has deduplication, 579 00:22:46,250 --> 00:22:50,000 it has enrichments from CSV filtering. 580 00:22:51,875 --> 00:22:54,958 And I will show one specific use case 581 00:22:54,958 --> 00:22:57,208 which is covering workflows. 582 00:22:57,833 --> 00:23:00,916 And I will try launching the whole 583 00:23:00,916 --> 00:23:02,375 software from ground up 584 00:23:02,375 --> 00:23:04,666 as typical open source user 585 00:23:06,458 --> 00:23:08,666 to bring more risks to this demo. 586 00:23:09,500 --> 00:23:09,666 Right. 587 00:23:09,958 --> 00:23:10,208 I see. 588 00:23:10,500 --> 00:23:11,083 We love that. 589 00:23:11,625 --> 00:23:14,916 We love because this is closer to what 590 00:23:14,916 --> 00:23:15,958 anyone would experience 591 00:23:15,958 --> 00:23:16,791 when they would start. 592 00:23:17,500 --> 00:23:19,458 This is the real deal. 593 00:23:20,750 --> 00:23:21,083 Thank you. 594 00:23:21,958 --> 00:23:24,541 So let's think about our use case. 595 00:23:25,166 --> 00:23:30,291 Usually, based on my experience, when 596 00:23:30,291 --> 00:23:33,250 people bring Keep into their 597 00:23:33,250 --> 00:23:34,708 infrastructure, they 598 00:23:34,708 --> 00:23:38,333 want as a step one for them is to 599 00:23:38,333 --> 00:23:40,375 integrate it well with 600 00:23:40,375 --> 00:23:41,666 everything else they have. 601 00:23:42,791 --> 00:23:45,750 And the exact ways they want to build 602 00:23:45,750 --> 00:23:48,958 this integration is always different. 603 00:23:50,041 --> 00:23:52,250 And I was thinking about some use case 604 00:23:52,250 --> 00:23:55,250 and the use case I want to cover today 605 00:23:55,583 --> 00:23:57,375 is receiving alerts 606 00:23:57,375 --> 00:23:58,791 from monitoring system. 607 00:23:59,625 --> 00:24:03,291 Once I received an alert, I want to go to 608 00:24:03,291 --> 00:24:04,375 actually application 609 00:24:04,375 --> 00:24:06,708 database and check something. 610 00:24:07,875 --> 00:24:12,083 And if, for example, my job didn't 611 00:24:12,083 --> 00:24:14,000 publish a result to the 612 00:24:14,000 --> 00:24:16,875 database in the last 15 minutes, 613 00:24:17,708 --> 00:24:19,958 I want to shoot this alert 614 00:24:19,958 --> 00:24:21,458 to the on call engineer 615 00:24:21,458 --> 00:24:22,833 to some IRM system. 616 00:24:23,416 --> 00:24:24,916 So I will build this 617 00:24:24,916 --> 00:24:26,208 use case from ground up. 618 00:24:26,666 --> 00:24:27,291 How does it sound? 619 00:24:27,916 --> 00:24:28,833 Yeah, sounds great. 620 00:24:29,166 --> 00:24:29,833 Let's try it! 621 00:24:34,083 --> 00:24:36,500 To see the demo, find the YouTube video 622 00:24:36,500 --> 00:24:37,958 link in the show notes. 623 00:24:38,708 --> 00:24:40,041 And if you enjoy this content 624 00:24:40,041 --> 00:24:42,541 and want to support it, go to 625 00:24:42,833 --> 00:24:46,083 makeitwork.tv, join as a member 626 00:24:46,083 --> 00:24:47,375 and watch the full 627 00:24:47,375 --> 00:24:50,083 conversation in 4K straight off the 628 00:24:50,083 --> 00:24:51,458 Jellyfin media server. 629 00:24:52,250 --> 00:24:53,916 Yes, offline download 630 00:24:54,083 --> 00:24:54,750 is enabled. 631 00:24:58,541 --> 00:25:00,333 We now rejoin Matvey and Tal 632 00:25:00,333 --> 00:25:02,375 just as they made it work. 633 00:25:05,041 --> 00:25:06,083 Moment of truth. 634 00:25:09,333 --> 00:25:09,958 Okay 635 00:25:10,541 --> 00:25:11,041 It's spinning. 636 00:25:11,041 --> 00:25:12,750 Wow, 200. 637 00:25:14,208 --> 00:25:15,458 And we received our 638 00:25:15,458 --> 00:25:16,333 alert with environment. 639 00:25:17,125 --> 00:25:18,166 Thank you Tal for the help. 640 00:25:18,791 --> 00:25:19,291 It worked. 641 00:25:21,958 --> 00:25:23,416 And that's it. 642 00:25:23,625 --> 00:25:24,666 Let's sum it up. 643 00:25:25,208 --> 00:25:26,541 So what did we do? 644 00:25:26,625 --> 00:25:27,958 We received an alert 645 00:25:27,958 --> 00:25:29,916 from the monitoring system. 646 00:25:31,541 --> 00:25:32,708 We went to... 647 00:25:33,208 --> 00:25:35,500 We built a workflow which is automatically 648 00:25:35,791 --> 00:25:37,125 going to the MySQL 649 00:25:37,125 --> 00:25:38,833 database, making a query. 650 00:25:39,458 --> 00:25:41,208 And based on the results of this query, 651 00:25:41,250 --> 00:25:42,916 it's republishing this 652 00:25:42,916 --> 00:25:44,625 alert to a third party system. 653 00:25:47,458 --> 00:25:48,500 That's actually it. 654 00:25:48,833 --> 00:25:49,375 OK. 655 00:25:49,750 --> 00:25:54,500 So does this need an alert 656 00:25:55,291 --> 00:25:57,125 to be triggered for 657 00:25:57,125 --> 00:25:58,291 the workflow to start? 658 00:25:59,000 --> 00:26:00,250 So I know that we mentioned that 659 00:26:00,250 --> 00:26:01,791 workflows can be triggered 660 00:26:01,791 --> 00:26:03,000 by different types of events. 661 00:26:03,750 --> 00:26:05,541 We mentioned an incident 662 00:26:05,541 --> 00:26:07,083 can trigger a workflow. 663 00:26:07,833 --> 00:26:10,166 We've seen an alert triggering a workflow. 664 00:26:11,000 --> 00:26:12,208 Is there something else 665 00:26:12,208 --> 00:26:13,708 that can trigger workflows? 666 00:26:15,000 --> 00:26:16,333 There is one more option 667 00:26:16,333 --> 00:26:18,375 to trigger them manually. 668 00:26:20,166 --> 00:26:22,166 But if I trigger them 669 00:26:22,166 --> 00:26:23,458 manually, I will not 670 00:26:23,500 --> 00:26:26,125 have alert content here. 671 00:26:27,125 --> 00:26:28,958 So I intentionally, when I was debugging, 672 00:26:29,166 --> 00:26:29,833 I wanted to... 673 00:26:32,458 --> 00:26:33,541 to debug it using alert. 674 00:26:34,458 --> 00:26:35,250 Probably something for 675 00:26:35,250 --> 00:26:37,041 us to improve in the UI. 676 00:26:38,708 --> 00:26:42,708 You can run it manually for the alert. 677 00:26:43,166 --> 00:26:43,833 Right. 678 00:26:44,583 --> 00:26:46,333 Yeah, it's already implemented, yeah. 679 00:26:46,958 --> 00:26:50,416 One of the things we had in mind here 680 00:26:50,500 --> 00:26:53,250 is that one-- another 681 00:26:53,250 --> 00:26:54,625 pain point with alerts 682 00:26:54,875 --> 00:26:56,416 is that you always find 683 00:26:56,416 --> 00:26:58,583 it hard to debug them, 684 00:26:58,583 --> 00:26:59,500 to understand whether 685 00:26:59,500 --> 00:27:00,708 it really works or not. 686 00:27:01,250 --> 00:27:03,750 Like just what Matvey just did. 687 00:27:04,083 --> 00:27:06,833 So sending the alert over and over again, 688 00:27:06,875 --> 00:27:08,833 see if the threshold is right. 689 00:27:08,833 --> 00:27:10,000 Maybe you need to adjust it. 690 00:27:10,000 --> 00:27:11,416 Maybe you need to change something. 691 00:27:11,458 --> 00:27:13,916 Maybe there is some typo in 692 00:27:13,916 --> 00:27:16,208 the query you just did. 693 00:27:17,375 --> 00:27:18,666 And this was one of the things we 694 00:27:18,666 --> 00:27:20,541 wanted to implement within 695 00:27:20,541 --> 00:27:22,166 Keep, is to allow somebody 696 00:27:22,333 --> 00:27:25,375 to test this workflow very easily. 697 00:27:25,583 --> 00:27:27,833 So instead of having to send the 698 00:27:27,833 --> 00:27:29,708 triggering event over and over 699 00:27:29,791 --> 00:27:31,166 again, you can just very 700 00:27:31,166 --> 00:27:33,708 simply either manually execute 701 00:27:34,041 --> 00:27:36,416 the workflow and feed it 702 00:27:36,416 --> 00:27:38,291 with everything that it expects. 703 00:27:38,791 --> 00:27:41,208 So for example, environment in this case, 704 00:27:41,291 --> 00:27:42,916 you could just fill it in manually 705 00:27:42,916 --> 00:27:46,250 and test if the workflow runs properly. 706 00:27:46,416 --> 00:27:48,500 Or you can actually take a past alert 707 00:27:48,708 --> 00:27:52,750 and make it trigger the workflow again 708 00:27:52,750 --> 00:27:54,916 and fill in all the information. 709 00:27:56,375 --> 00:27:57,625 One more thing that I 710 00:27:57,625 --> 00:27:58,833 think is interesting 711 00:27:59,041 --> 00:28:03,166 is that besides events, you can also 712 00:28:03,291 --> 00:28:05,416 have interval, which is 713 00:28:05,416 --> 00:28:09,458 basically also a very common use 714 00:28:09,583 --> 00:28:11,208 case that we see from our users. 715 00:28:11,583 --> 00:28:13,916 Just run this workflow every 30 seconds 716 00:28:14,125 --> 00:28:17,583 to check if the event happened or not, 717 00:28:17,750 --> 00:28:20,708 and then decide to do something or not. 718 00:28:20,750 --> 00:28:22,500 Do you have any sort 719 00:28:22,500 --> 00:28:23,875 of reporting capability 720 00:28:24,666 --> 00:28:26,291 in terms of showing how 721 00:28:26,291 --> 00:28:27,583 many alerts were triggered, 722 00:28:28,416 --> 00:28:29,083 workflows? 723 00:28:29,833 --> 00:28:30,958 And I'm also wondering, 724 00:28:30,958 --> 00:28:32,125 because off of this, 725 00:28:32,750 --> 00:28:34,000 we've seen workflows fail. 726 00:28:34,708 --> 00:28:35,500 So would you trigger an 727 00:28:35,500 --> 00:28:36,875 alert when the workflow fails? 728 00:28:38,416 --> 00:28:41,791 I think it's a cool feature idea. 729 00:28:42,125 --> 00:28:43,083 We actually have emails 730 00:28:43,083 --> 00:28:45,500 sending when workflows fail. 731 00:28:45,833 --> 00:28:47,416 So if Matvey is the one 732 00:28:47,416 --> 00:28:48,916 who uploaded the workflow, 733 00:28:48,916 --> 00:28:51,000 he will get an email saying that the 734 00:28:51,000 --> 00:28:52,916 execution of the workflow failed. 735 00:28:54,208 --> 00:28:56,416 But now that you mentioned it, I think 736 00:28:56,458 --> 00:28:57,791 it's super cool to 737 00:28:57,791 --> 00:28:59,250 maybe have a workflow that 738 00:28:59,250 --> 00:29:01,916 runs when other workflows are failing. 739 00:29:02,333 --> 00:29:03,916 That's a cool idea. 740 00:29:04,541 --> 00:29:07,125 Yeah, because that would be like my worry 741 00:29:07,250 --> 00:29:08,750 when I set the system like this up. 742 00:29:09,208 --> 00:29:10,416 What monitors the system 743 00:29:10,416 --> 00:29:11,666 that's supposed to alert me? 744 00:29:12,083 --> 00:29:13,125 And when there's a failure, 745 00:29:13,125 --> 00:29:14,541 I would want to know about it. 746 00:29:15,458 --> 00:29:17,541 I mean, the fact that you 747 00:29:17,541 --> 00:29:18,791 have emails is a good idea, 748 00:29:19,250 --> 00:29:20,041 because maybe you shouldn't 749 00:29:20,041 --> 00:29:21,583 use the same thing that maybe is 750 00:29:21,583 --> 00:29:23,541 failing, because then you end up in a 751 00:29:23,541 --> 00:29:25,166 constant endless loop. 752 00:29:25,541 --> 00:29:26,333 The thing is failing. 753 00:29:26,333 --> 00:29:27,583 It's triggering other workflows. 754 00:29:27,583 --> 00:29:28,166 Maybe there's like-- 755 00:29:28,416 --> 00:29:30,291 I can imagine that being a problem. 756 00:29:30,708 --> 00:29:33,208 So having emails is important, again, 757 00:29:33,208 --> 00:29:33,875 so that it gives you 758 00:29:33,875 --> 00:29:35,250 assurance that when there's a problem, 759 00:29:35,458 --> 00:29:36,333 you will know about it. 760 00:29:36,791 --> 00:29:37,958 Actually, it's a really 761 00:29:37,958 --> 00:29:39,875 good question about metrics 762 00:29:40,250 --> 00:29:43,541 and about observability for 763 00:29:43,541 --> 00:29:45,208 Keep, like how Keep works. 764 00:29:46,250 --> 00:29:47,500 And that's something we 765 00:29:47,500 --> 00:29:49,791 recently started to invest more. 766 00:29:50,541 --> 00:29:51,875 And one of the latest 767 00:29:51,875 --> 00:29:54,041 features is a metrics endpoint, 768 00:29:54,375 --> 00:29:55,208 actually, for primitives. 769 00:29:55,875 --> 00:29:58,250 So you could export to Prometheus 770 00:29:59,666 --> 00:30:01,625 how many alerts you have, 771 00:30:02,583 --> 00:30:04,333 how many alerts per incident you 772 00:30:04,416 --> 00:30:08,458 have, and actually, you 773 00:30:08,458 --> 00:30:11,791 could filter by labels here 774 00:30:11,791 --> 00:30:12,583 from alerts. 775 00:30:12,750 --> 00:30:14,750 So this metric 776 00:30:14,750 --> 00:30:17,000 endpoint, it seems like small, 777 00:30:17,000 --> 00:30:18,458 but it's really powerful in order 778 00:30:18,458 --> 00:30:22,958 to get some usage statistics back to you. 779 00:30:23,666 --> 00:30:24,000 Yeah. 780 00:30:25,041 --> 00:30:27,000 So when it comes to deploying Keep, 781 00:30:27,375 --> 00:30:28,125 we've seen like the 782 00:30:28,125 --> 00:30:29,333 Docker Compose approach. 783 00:30:29,375 --> 00:30:30,625 We've also seen that there 784 00:30:30,625 --> 00:30:32,333 is a way to deploy it in Kubernetes. 785 00:30:33,208 --> 00:30:34,291 I'm wondering what are the 786 00:30:34,291 --> 00:30:36,791 requirements for Keep to run. 787 00:30:37,166 --> 00:30:38,000 So it's sending emails. 788 00:30:38,291 --> 00:30:39,291 I'm assuming there needs to 789 00:30:39,291 --> 00:30:40,208 be some sort of integration 790 00:30:40,208 --> 00:30:43,250 with like an email provider of some sort, 791 00:30:43,416 --> 00:30:45,291 to be able to send those emails, maybe 792 00:30:45,291 --> 00:30:48,833 SMTP, but also a database. 793 00:30:49,333 --> 00:30:50,458 So does it use SQLite? 794 00:30:50,708 --> 00:30:51,541 Do you need MySQL? 795 00:30:52,291 --> 00:30:55,166 How does the Keep installation, like the 796 00:30:55,166 --> 00:30:56,166 simplest Keep installation, 797 00:30:56,166 --> 00:30:56,833 look like? 798 00:30:58,375 --> 00:31:00,125 So yeah, we have a few 799 00:31:00,125 --> 00:31:02,083 different deployment types. 800 00:31:03,333 --> 00:31:04,916 Actually, Matvey, maybe you 801 00:31:04,916 --> 00:31:07,041 can open the specification page 802 00:31:07,458 --> 00:31:09,125 from the documentation once again. 803 00:31:09,125 --> 00:31:12,875 But we have-- I'm just 804 00:31:12,875 --> 00:31:15,541 discussing it with our team 805 00:31:15,541 --> 00:31:16,541 right now, but I think 806 00:31:16,541 --> 00:31:19,541 it's somewhat shooting 807 00:31:19,541 --> 00:31:21,666 ourself in the foot. 808 00:31:22,250 --> 00:31:23,625 But we support various 809 00:31:23,625 --> 00:31:26,541 different dialects of databases. 810 00:31:26,916 --> 00:31:32,875 So both SQLite, MySQL, Postgres, and 811 00:31:32,875 --> 00:31:34,541 Microsoft SQL Server. 812 00:31:35,291 --> 00:31:37,291 Some of our customers have that. 813 00:31:38,625 --> 00:31:40,541 It's an engineering 814 00:31:40,541 --> 00:31:42,958 experience to do that. 815 00:31:43,166 --> 00:31:44,541 But this is what we support. 816 00:31:44,958 --> 00:31:46,500 It very much depends on 817 00:31:46,500 --> 00:31:48,416 the scale of the events 818 00:31:48,416 --> 00:31:49,666 that you're sending into Keep. 819 00:31:50,958 --> 00:31:55,875 So we have a page that describes what 820 00:31:55,875 --> 00:31:57,750 kind of infrastructure 821 00:31:58,000 --> 00:32:02,500 do we expect in different scale settings. 822 00:32:04,000 --> 00:32:05,125 Something else we have 823 00:32:05,125 --> 00:32:06,583 within the infrastructure 824 00:32:06,875 --> 00:32:09,500 is at some scale, you need to 825 00:32:09,500 --> 00:32:11,333 also have some queuing system 826 00:32:11,333 --> 00:32:13,791 for the events to properly digest. 827 00:32:14,583 --> 00:32:16,791 So we also use Redis with something 828 00:32:16,791 --> 00:32:19,000 that is called ARQ, which is basically 829 00:32:19,333 --> 00:32:23,666 an implementation for queues over Redis. 830 00:32:25,625 --> 00:32:27,541 But this page in our documentation 831 00:32:27,875 --> 00:32:29,791 basically describes everything that we 832 00:32:29,791 --> 00:32:31,666 need in terms of CPU 833 00:32:31,666 --> 00:32:32,333 and memory. 834 00:32:33,541 --> 00:32:34,541 And we support, again, 835 00:32:35,583 --> 00:32:36,875 Kubernetes, OpenShift. 836 00:32:38,208 --> 00:32:41,708 Somebody from our community wrote a guide 837 00:32:41,708 --> 00:32:43,375 on how to deploy this to 838 00:32:43,375 --> 00:32:46,708 AWS ECS, which is also cool. 839 00:32:48,541 --> 00:32:50,500 OK, this is great. 840 00:32:51,750 --> 00:32:53,208 That was a cool demo, by the way, Matvey. 841 00:32:53,208 --> 00:32:54,041 Thank you very much. 842 00:32:54,416 --> 00:32:55,125 Oh, thank you. 843 00:32:55,208 --> 00:32:55,708 And the fact that it didn't 844 00:32:55,708 --> 00:32:57,625 fully work the first time, that's how we 845 00:32:57,625 --> 00:32:58,875 know it was real, right? 846 00:32:58,875 --> 00:33:00,583 It wasn't recorded, right? 847 00:33:00,750 --> 00:33:02,791 We did it, you know, we insert like all 848 00:33:02,791 --> 00:33:04,500 the commands and we started from scratch. 849 00:33:04,500 --> 00:33:05,916 And there was like a small little thing 850 00:33:05,916 --> 00:33:07,250 that we had to figure out, but it was 851 00:33:07,250 --> 00:33:09,416 all very smooth, I have to say. 852 00:33:09,416 --> 00:33:10,833 So thank you very much for sharing that. 853 00:33:11,625 --> 00:33:12,000 Thank you. 854 00:33:12,500 --> 00:33:16,208 As we are approaching the end, 855 00:33:16,208 --> 00:33:19,916 I'm wondering what are you thinking for next 856 00:33:19,916 --> 00:33:22,166 year in the context of Keep? 857 00:33:23,125 --> 00:33:24,916 Are there some things that you would like 858 00:33:24,916 --> 00:33:27,125 to get to, some maybe big 859 00:33:27,125 --> 00:33:28,500 challenges that you see ahead? 860 00:33:28,875 --> 00:33:32,000 How do you see 2025 for Keep? 861 00:33:33,166 --> 00:33:37,333 I can share my thoughts about this and 862 00:33:37,333 --> 00:33:40,375 it's generally two 863 00:33:40,375 --> 00:33:42,958 paths that I see for Keep. 864 00:33:43,958 --> 00:33:45,625 So something like we didn't really 865 00:33:45,625 --> 00:33:50,166 discuss but it's a major part of what 866 00:33:50,166 --> 00:33:52,041 Keep is, is that we're open source. 867 00:33:53,041 --> 00:33:55,416 And like we discussed before, there's a 868 00:33:55,416 --> 00:33:57,291 lot of use cases for the smaller 869 00:33:57,291 --> 00:33:58,041 startups, smaller 870 00:33:58,041 --> 00:33:59,750 companies when they use Keep. 871 00:34:00,458 --> 00:34:05,458 And I think we love open source besides 872 00:34:05,458 --> 00:34:08,541 the fact that it opens a business world 873 00:34:08,541 --> 00:34:10,458 for us because also 874 00:34:10,458 --> 00:34:12,125 bigger companies, enterprises, 875 00:34:12,458 --> 00:34:16,125 they look at open source as the big next 876 00:34:16,125 --> 00:34:19,500 thing and they try to migrate some of the 877 00:34:19,500 --> 00:34:20,791 tools that they're already 878 00:34:20,791 --> 00:34:22,458 using today to open source. 879 00:34:22,458 --> 00:34:28,708 So one path in my eyes is that we want to 880 00:34:28,708 --> 00:34:30,333 keep on nurturing this community. 881 00:34:31,000 --> 00:34:33,541 We want to make it, we want to make Keep 882 00:34:33,541 --> 00:34:38,583 the go-to for engineers or for operations 883 00:34:38,583 --> 00:34:43,083 as this Swiss knife of everything that is 884 00:34:43,083 --> 00:34:45,333 alerts or events and the 885 00:34:45,333 --> 00:34:46,458 things they can do with it. 886 00:34:47,458 --> 00:34:50,583 On the other hand, we have huge, huge, 887 00:34:50,583 --> 00:34:55,625 huge things that we still need to figure 888 00:34:55,625 --> 00:34:58,458 out with AI and everything that is 889 00:34:58,458 --> 00:35:02,333 evolving practically every minute. 890 00:35:02,625 --> 00:35:05,583 Like I guess there's a new GPT model that 891 00:35:05,583 --> 00:35:06,916 just released while we 892 00:35:06,916 --> 00:35:09,750 were recording this podcast. 893 00:35:10,458 --> 00:35:13,458 So there's a lot of things we need to 894 00:35:13,458 --> 00:35:15,833 figure out there. There's a lot of work. 895 00:35:17,250 --> 00:35:19,833 One of the things you mentioned a few 896 00:35:19,833 --> 00:35:24,291 minutes ago without maybe even knowing 897 00:35:24,291 --> 00:35:25,791 how important it is for 898 00:35:25,791 --> 00:35:28,000 us is the reliability part. 899 00:35:28,000 --> 00:35:30,333 So how can we make sure we are reliable 900 00:35:30,333 --> 00:35:33,916 being the most crucial part in the 901 00:35:33,916 --> 00:35:37,208 reliability pipeline in companies? 902 00:35:37,458 --> 00:35:39,416 And that's something we need to invest a 903 00:35:39,416 --> 00:35:43,458 lot in. And I think that's the two major, 904 00:35:43,458 --> 00:35:45,500 major things we are 905 00:35:45,500 --> 00:35:47,916 focusing on right now. 906 00:35:48,250 --> 00:35:50,083 Matvey, I don't know if you see the 907 00:35:50,083 --> 00:35:53,708 things the same way, but I guess I can 908 00:35:53,708 --> 00:35:56,166 guess you can agree with that. 909 00:35:57,083 --> 00:35:57,291 Absolutely. 910 00:35:58,958 --> 00:36:04,000 Now you hash it out. 911 00:36:05,458 --> 00:36:09,416 I am like we are very, very much an 912 00:36:09,416 --> 00:36:12,208 engineering team and all of us are 913 00:36:12,208 --> 00:36:13,416 engineering founders. 914 00:36:13,958 --> 00:36:15,750 And when we speak about technical 915 00:36:15,750 --> 00:36:18,666 challenges like reliability, I'm not 916 00:36:18,666 --> 00:36:20,416 worried about them because technical 917 00:36:20,416 --> 00:36:22,458 challenges are not that challenging here. 918 00:36:23,458 --> 00:36:28,708 What's challenging actually is to prove 919 00:36:28,708 --> 00:36:31,250 the market that we can do what we claim 920 00:36:31,250 --> 00:36:34,166 in the AI space, which we almost didn't 921 00:36:34,166 --> 00:36:38,833 touch today, that we can correlate alerts 922 00:36:38,833 --> 00:36:43,541 to incidents at scale on premises 923 00:36:43,541 --> 00:36:46,833 environment with no internet connection. 924 00:36:47,500 --> 00:36:50,791 We could summarize incidents, provide 925 00:36:50,791 --> 00:36:54,458 insights from the past runbooks and 926 00:36:54,458 --> 00:36:56,375 that's a huge scope of work we do. 927 00:36:57,041 --> 00:36:59,333 And that's actually what we sell. And we 928 00:36:59,333 --> 00:37:03,583 see that in the market nowadays, there is 929 00:37:03,583 --> 00:37:06,458 a lot of skepticism about it. 930 00:37:07,458 --> 00:37:11,041 Like there are few products who tried to 931 00:37:11,041 --> 00:37:13,208 do it in the past and 932 00:37:13,208 --> 00:37:14,958 didn't do it that well. 933 00:37:16,083 --> 00:37:20,958 And people who tried them, they're like, 934 00:37:21,250 --> 00:37:23,416 maybe it's not a real, real thing. 935 00:37:24,458 --> 00:37:27,708 Who tried it like many years ago, it 936 00:37:27,708 --> 00:37:29,458 didn't work that well actually before. 937 00:37:30,458 --> 00:37:33,333 But nowadays with what we see in the AI 938 00:37:33,333 --> 00:37:37,125 world with LLMs, with local models, open 939 00:37:37,125 --> 00:37:40,458 source models, it's all getting possible. 940 00:37:41,458 --> 00:37:46,041 And challenge for us, huge challenge for 941 00:37:46,041 --> 00:37:51,791 us for 2025 is to have like a few good 942 00:37:51,791 --> 00:37:54,375 examples how it works at scale for large 943 00:37:54,375 --> 00:37:57,166 customers and make them 944 00:37:57,166 --> 00:38:01,500 published with customer stories to change 945 00:38:01,500 --> 00:38:05,750 the perception of AI ops in general, that 946 00:38:05,750 --> 00:38:08,000 AI ops could actually be AI. 947 00:38:08,500 --> 00:38:12,041 Right. If people want to learn more about 948 00:38:12,041 --> 00:38:14,333 this side of Keep, where should they go 949 00:38:14,333 --> 00:38:15,833 to read or watch 950 00:38:15,833 --> 00:38:17,250 something? What would you recommend? 951 00:38:18,416 --> 00:38:20,500 We will publish more soon. 952 00:38:21,166 --> 00:38:23,041 For now, the best way 953 00:38:23,041 --> 00:38:25,000 is to go to our website. 954 00:38:25,291 --> 00:38:27,166 There is a contact us form, and we 955 00:38:27,166 --> 00:38:28,875 actually just do demos. 956 00:38:29,500 --> 00:38:30,583 That's the best way. 957 00:38:31,166 --> 00:38:32,708 The very similar call, but 958 00:38:32,708 --> 00:38:34,958 mostly focused on the AI part, 959 00:38:35,000 --> 00:38:36,666 we run a huge instance, 960 00:38:37,458 --> 00:38:38,541 which with a lot of alerts 961 00:38:38,583 --> 00:38:40,375 and it's happening on the 962 00:38:40,375 --> 00:38:43,083 fly, sometimes doesn't work. 963 00:38:44,125 --> 00:38:44,500 Like today. 964 00:38:45,541 --> 00:38:47,125 But that's why we have demos. 965 00:38:49,125 --> 00:38:49,333 Cool. 966 00:38:50,250 --> 00:38:51,750 Matvey, thank you very much for the demo, 967 00:38:51,958 --> 00:38:53,791 and thank you very much for reconnecting 968 00:38:53,791 --> 00:38:55,333 after a couple of years. 969 00:38:55,541 --> 00:38:56,500 I really enjoyed that. 970 00:38:56,833 --> 00:38:59,416 Tal, it's been great knowing you and 971 00:38:59,416 --> 00:39:00,500 talking a little bit 972 00:39:00,916 --> 00:39:02,083 about what you're doing next. 973 00:39:02,625 --> 00:39:04,166 I'm excited to see what you do. 974 00:39:04,166 --> 00:39:04,500 You're right. 975 00:39:04,500 --> 00:39:05,125 It is challenging. 976 00:39:05,833 --> 00:39:07,416 The reliability part is challenging. 977 00:39:07,625 --> 00:39:09,083 I deal with that on a daily basis, 978 00:39:09,291 --> 00:39:10,375 and I know how far it goes. 979 00:39:11,291 --> 00:39:12,291 It's all systems, and 980 00:39:12,291 --> 00:39:13,666 people are part of those systems. 981 00:39:14,375 --> 00:39:16,416 So even though tech is the good part, 982 00:39:16,458 --> 00:39:17,666 it's the people that make mistakes. 983 00:39:18,125 --> 00:39:18,708 What you're trying to 984 00:39:18,708 --> 00:39:20,208 prevent, or at least we're 985 00:39:20,208 --> 00:39:21,416 trying to make it more 986 00:39:21,416 --> 00:39:22,291 obvious when they happen. 987 00:39:22,291 --> 00:39:23,416 So that's the hard part. 988 00:39:23,791 --> 00:39:25,291 So I do have an appreciation for that. 989 00:39:25,875 --> 00:39:28,541 But the AI, all the stuff 990 00:39:28,541 --> 00:39:29,541 that's coming, you're right. 991 00:39:29,541 --> 00:39:31,208 It's moving very fast, but the 992 00:39:31,208 --> 00:39:32,708 capabilities are amazing. 993 00:39:33,541 --> 00:39:34,666 So I'm excited about the 994 00:39:34,666 --> 00:39:36,166 intersection of the reliability 995 00:39:36,416 --> 00:39:39,333 part, the AI part, the complexity part, 996 00:39:39,333 --> 00:39:40,500 how we're trying to just 997 00:39:40,500 --> 00:39:42,166 make sense of the sea of alerts, 998 00:39:42,166 --> 00:39:42,916 which I think is a 999 00:39:42,916 --> 00:39:44,583 great way to think of it. 1000 00:39:45,000 --> 00:39:45,666 We're just getting 1001 00:39:45,666 --> 00:39:47,583 overloaded with signals 1002 00:39:47,583 --> 00:39:48,458 from all over the place. 1003 00:39:48,666 --> 00:39:49,500 Most of it is noise, 1004 00:39:50,250 --> 00:39:51,666 but in that haystack, 1005 00:39:51,666 --> 00:39:53,250 there's a needle that's really important. 1006 00:39:53,833 --> 00:39:55,375 And if you don't pay attention to it, 1007 00:39:55,958 --> 00:39:58,125 CrowdStrike 2 is just around the corner. 1008 00:39:59,000 --> 00:40:00,000 That's a placeholder for 1009 00:40:00,000 --> 00:40:01,458 whatever the next big thing is, 1010 00:40:01,458 --> 00:40:02,666 but we know it's going to happen. 1011 00:40:03,250 --> 00:40:04,375 Thank you for having us. 1012 00:40:05,000 --> 00:40:05,708 This was amazing. 1013 00:40:06,666 --> 00:40:09,041 And probably a sneak peek 1014 00:40:09,041 --> 00:40:12,083 for maybe we meet again. 1015 00:40:12,875 --> 00:40:13,791 I hope not in three 1016 00:40:13,791 --> 00:40:15,166 years, but maybe sooner. 1017 00:40:15,458 --> 00:40:15,666 Yes. 1018 00:40:15,958 --> 00:40:16,583 It is a teaser. 1019 00:40:17,000 --> 00:40:17,583 Sooner, yes. 1020 00:40:18,875 --> 00:40:20,416 We have some plans. 1021 00:40:21,291 --> 00:40:22,208 We spoke about the AI 1022 00:40:22,208 --> 00:40:24,333 today, and we intentionally 1023 00:40:24,583 --> 00:40:27,333 didn't show it because we find it 1024 00:40:27,666 --> 00:40:29,000 valuable to show open 1025 00:40:29,000 --> 00:40:31,208 source part in such podcasts, 1026 00:40:31,208 --> 00:40:32,541 something you can jump in. 1027 00:40:32,875 --> 00:40:33,083 Indeed. 1028 00:40:33,666 --> 00:40:35,083 And if you're working 1029 00:40:35,083 --> 00:40:36,000 in large enterprise, 1030 00:40:36,000 --> 00:40:36,708 you will reach out to 1031 00:40:36,708 --> 00:40:37,875 us and we'll show you AI. 1032 00:40:38,625 --> 00:40:40,041 But a sneak peek that 1033 00:40:40,041 --> 00:40:41,333 follow us because we 1034 00:40:41,375 --> 00:40:44,458 have a strong belief in open source, 1035 00:40:44,708 --> 00:40:45,791 and we really, really want 1036 00:40:45,791 --> 00:40:48,583 to downstream some of our AI 1037 00:40:48,625 --> 00:40:49,916 features to open source once 1038 00:40:49,916 --> 00:40:51,041 they will be polished enough. 1039 00:40:52,416 --> 00:40:53,291 So stay tuned. 1040 00:40:54,500 --> 00:40:54,833 Love that. 1041 00:40:55,666 --> 00:40:55,875 All right. 1042 00:40:56,166 --> 00:40:56,916 See you next time. 1043 00:40:57,041 --> 00:40:57,500 Bye bye. 1044 00:40:57,750 --> 00:40:58,166 Thank you. 1045 00:41:00,916 --> 00:41:02,791 If you enjoyed this episode, I will 1046 00:41:02,791 --> 00:41:04,791 appreciate a rate and review in your 1047 00:41:04,791 --> 00:41:06,166 favorite podcast app. 1048 00:41:06,958 --> 00:41:09,291 You may want to subscribe for free to 1049 00:41:09,291 --> 00:41:12,041 makeitwork.tv so that you don't miss 1050 00:41:12,041 --> 00:41:14,416 the one year anniversary newsletter post. 1051 00:41:15,083 --> 00:41:17,666 Thank you for tuning in, until next time.