0:10 So, my name is Muhammad Fil. Today is
0:13 August 13th, uh 10:36 a.m. Eastern
0:16 Standard Time. Uh I'm the chief examiner
0:19 of Anafi. So today we will be going
0:23 through um AI ops diploma artificial
0:26 intelligence operations uh RQF level six
0:31 educall level six exam uh the candidate
0:34 name is Abdullah Sal
0:37 uh the topic is pretty uh interesting uh
0:39 implementing autonomous site reliability
0:42 engineering with uh captain Argo
0:44 rollouts and Prometheus for e-commerce
0:47 right uh site reliability engineers is
0:50 are in high demand right now, right?
0:53 Organizations are looking for these
0:55 kinds of people
1:00 like they're like know I would say they
1:03 are paid very well when it comes to u
1:06 engineering disciplines nowadays as
1:07 computer system administrators and
1:12 societal right. So this is a very
1:14 detailed topic Abdullah.
1:16 Uh the key learning objectives are
1:18 autonomous deployment and progressive
1:21 delivery, advanced observability in SLO
1:24 service level uh management,
1:26 uh intelligent autoscaling and resource
1:29 optimization, automated in instant
1:31 response and remediation, cost
1:33 optimization, resource management,
1:36 performance testing and load management
1:39 and industry applications. Right? topic
1:42 is very loaded. Um you'll be given 15 to
1:43 20 minutes for the presentation followed
1:47 by 30 to 40 minutes for in
1:48 are you ready? >> Yeah.
1:49 >> Yeah.
1:51 >> Do you have any questions before we start?
1:52 start?
1:53 >> Uh no sir.
2:09 There is an echo sound or like a white
2:12 noise sound coming from your background.
2:15 I guess that's okay.
2:16 >> That might that might be the fan. That
2:18 might be something else.
2:19 >> Uh yeah, it's f.
2:22 >> Yes. Right. It says continuous white noise.
2:23 noise. >> Yeah,
2:24 >> Yeah,
2:26 >> you can turn it off. That would be good.
2:29 If not, we can we can proceed.
2:32 >> So I will try.
3:15 This is now better. It's better than the
3:18 past. Okay. All right. So,
3:22 you may start your presentation now.
3:25 >> Okay. Uh, good evening everyone. My name
3:26 is Abdullah.
3:29 Um, I'm from Saroda, Pakistan. I'm
3:32 enrolled in educ level six exam.
3:34 Today I will present my project
3:36 implementing autonomous site reliable
3:39 engineering with captain or rollout and
3:43 prometheus for our e-commerce scale.
3:45 This solution is all about modernizing
3:48 how we manage reliability, performance
3:51 and scaling in large scale e-commerce
3:55 system using intelligent automation
3:57 and self-healing deployment especially
3:59 for those applications who suffer from
4:02 slowdown system crash due to
4:06 unpredictable traffic companies like
4:08 Amazon, Shopify, Alibaba spending
4:16 Let's move to the agenda of our project.
4:20 Here is what we will cover. Our project
4:22 goal, challenges in our e-commerce
4:25 system, our system architecture and
4:27 detailed discussion on observability,
4:30 autoscaling, incident response and cost optimization
4:32 optimization
4:34 and more. Ending with real world
4:37 applications and key takeaways.
4:39 We will start with an overview of the
4:42 project then understand why autonomous s
4:45 matters. We will also take a look at uh
4:52 Let's move to the project overview. In
4:54 this project
4:56 uh we focus on creating a fully
4:58 automated self-managed reliability
5:02 system for cloudnative applications. It
5:05 integrate multiple tools like captain,
5:07 arco, prometheus
5:10 that handle things like safe deployment,
5:12 automatic recovery and system monitoring.
5:13 monitoring.
5:15 It's especially useful for e-commerce
5:18 platform that cannot afford a single
5:20 second of downtime.
5:22 On the coming slide, we will also
5:25 discuss about the tools we are using and
5:27 what are the purposes of those tools in
5:34 Let's start with a question.
5:36 Why autonomous S sur and how it is
5:40 better than traditional S sur? As we
5:42 discussed early uh our e-commerce
5:45 traffic is unpredictable and managing
5:47 everything manually can be timeconuming
5:50 and errorprone.
5:53 So uh autonomous s sur help us automate
5:56 reliability. So when the traffic spikes
5:58 or something breaks the system can
6:01 response automatically without need of a
6:04 human to jump in every time.
6:07 Uh this will improve uptime and user experience.
6:15 In this project, I have learned how to
6:18 build feedbackdriven system
6:21 using progressive deployment strategies,
6:24 monitoring system in real time, scale
6:27 application based on usage and business needs.
6:29 needs.
6:31 Performance testing and use uh
6:33 performance testing using different
6:37 tools uh before going live. optimization
6:40 cloud post using smart tools and we also
6:42 used some tools that responds uh to
6:50 Here is the architecture diagram that
6:53 shows everything uh how everything fit together.
6:55 together.
6:58 At the top we have deployment tools uh
7:00 which is uh like captain and argo rollouts.
7:02 rollouts.
7:04 In the second layer we have
7:07 observabilities tool uh Prometheus and
7:10 Garpon that help us monitor everything.
7:14 Autoscaler HPA and BPA that adjust the
7:17 resources based on CPU and memory
7:20 and if something goes wrong incident
7:23 response uh tools like packet duty and
7:27 NCL AWX take over. For cost
7:32 optimization, we have QOST and Goldilo.
7:35 Each part share data with the other uh
7:46 Autonomous deployment explained
7:49 instead of releasing all code at once
7:52 which is very risky and uh can also
7:55 cause a downtimes and system crash. So
7:57 we use safe deployment strategies like
8:00 Kendry and blue green deployment
8:02 tools like Argo rollout slowly shift
8:05 traffic uh while captain checks the
8:08 performance in real time that keep an
8:11 eye on SLOs's during deployment. If a
8:14 problem appear the rollout stop and roll
8:18 back this way uh bad code never reaches
8:21 all user and our system remain healthy
8:28 Deployment tools we uh there are three
8:31 key uh deployment tools we are using uh
8:33 which is Argo rollout, captain and
8:36 player. Argo rollouts control how
8:39 traffic is shifting during deployment
8:43 using canray and blue green deployment.
8:46 Captain uh act like a quality controller
8:47 checks if everything are running
8:50 smoothly and deployment is not violating
8:54 SLOs's like latency and error rate.
8:57 Tracker support this by controlling and
9:00 automating the rollout process uh
9:02 shifting traffic intelligently from old
9:05 to new uh from old to new version. Uh
9:08 they together make deployment safe and secure.
9:16 obser uh advanced observabilities we
9:18 need uh we need to know what happened
9:21 inside our system at all time. Advanced
9:25 observabilities mean we track trace and
9:27 simulate user behavior. This give us
9:30 early warning and deep visibility to fix
9:33 issues quickly even before the user
9:36 notice. It enable us to uh for faster
9:40 incident response uh for faster incident
9:43 response. The tools we are using for uh
9:46 observabilities are Prometheus and Garpona.
9:51 Prometheus collect all the performance
9:54 data from metrics. Garpona turns them
9:56 into a nice dashboard so we can easily
9:59 see what's happening.
10:02 Jagger and open telemetry shows how
10:05 traffic are talking uh how service are
10:07 talking with each other uh by checking
10:10 the uh request travel that shows uh
10:13 which part is a bottleneck uh which part
10:17 is a bottleneck and uh blackbox exporter
10:19 and synthetic monitoring uh sends fake
10:21 traffic to test availability of our
10:24 system. Combined they give us a complete picture.
10:33 Intelligent autoscaling traffic can
10:36 change fast. So our application need to
10:40 adjust on its own. Uh because in certain
10:43 traffic changes can cause downtime and
10:45 system crash. Autoscaling let us
10:48 increase or reduce resources depends on
10:52 how much load we are getting.
10:56 uh uh we use V HPA, VPA and KDA for
10:58 resource scaling based on uh based on
11:01 traffic patterns. This way we don't
11:10 There are the tools we are using uh with
11:13 uh mixed tools we are using uh like HPA
11:16 and VPA that handle port level scaling
11:19 uh that is based on CPU and memory while
11:22 ketta uh while ka react to business
11:25 events like uh uh like in black Friday
11:28 uh 111 and
11:31 uh and in some sales uh and in some sale events.
11:32 events.
11:35 Cluster autoscaler manage infrastructure
11:39 size increase or decrease nodes uh on in
11:41 our cluster
11:44 uh in our cluster together uh they keep
11:54 automated incident response if something
11:56 break we don't wait for uh someone to
11:59 fix it manually
12:01 uh if system detects the issue and kick
12:04 off a Pred uh if the system detects the
12:06 issue uh it uh kick off a predefined
12:10 issue uh predefined response
12:12 uh predefined respond that fix the
12:15 issues uh that fix the issues like uh uh
12:17 restarting restarting board scaling uh
12:22 resources or deleting cache caches.
12:24 uh this make recovery fast and reduce
12:34 incident response tools patcher duty
12:36 help us notify the right people and
12:40 trigger automation nibel AWX runs
12:42 playbooks that fix issues automatically
12:45 like restarting board and scaling
12:48 chaos monkey and germline
12:51 uh simulates user failure so we can test
13:00 cost optimization strategies
13:03 in the cloud. Using too many resources
13:06 cost money. So we use tools and
13:10 strategies to rights size our resources.
13:12 Move workload to cheaper machine and
13:16 constantly look for waste uh uh that can
13:18 uh that can remove without hurting the performance.
13:24 tools we are using for cost optimization.
13:26 optimization.
13:30 Qost show detail cost uh detail cost
13:32 breakdowns. It uh shows uh where we are
13:35 overpaying uh and if something uh is
13:38 expensive we will fix it.
13:41 Cool log tell us best CPU and memory settings.
13:43 settings.
13:46 Sport oceans and carp painter find uh
13:48 let us find the cheapest instance
13:51 available. uh all of these uh help
13:53 reduce billing without affecting reliability,
14:00 performance testing and load management.
14:03 Before user hit the system, we want to
14:05 know if it can handle the load. So we
14:08 run performance test of uh performance
14:12 test uh set up uh set up a load
14:14 performance uh load load protection like
14:16 circuit break uh and plan resources
14:26 Performance tools
14:29 tools like Ksix run load testing as a
14:32 part of our pipeline. K6 simulate user
14:36 traffic and perform uh and performance
14:39 testing during CI/CD pipeline. We also
14:42 use predictive models that learn from
14:44 the old traffic to tell us when to spikes
14:51 industrial application.
14:54 The architecture is not just theory. Big
14:56 companies like Amazon, Flipkart,
14:59 Alibaba, Das, eBay using the similar
15:03 setup. uh if uh it help us uh it helped
15:06 them to stay online and past uh even
15:07 millions of people visiting their
15:10 website at the same time
15:12 if uh it help us manage millions of
15:14 transaction during event like black
15:17 Friday, Cyber Monday 111
15:28 open source and cloud native tools
15:31 everything we use in Everything we use
15:33 is open source and cloudnative that
15:36 include Argo, Captain, Prometheus,
15:40 Karpana, Jagger, Ka and more. These
15:43 tools are reliable, well supported and
15:46 free to use making them great for setup,
15:56 This project bring together automation,
15:59 observability, scaling and performance
16:01 into a single platform that can run
16:04 itself. It shows how modern S sur can
16:07 eliminate manual work, reduce downtime
16:10 and save cost.
16:12 Thank you for listening. I'm happy to
16:46 continue. Abdullah.
16:54 Uh, it's the end of our my presentation.
16:58 >> End of the presentation. Okay.
17:00 >> Okay. Just let me compile the notes.
17:01 Hold on. >> Okay.
18:54 Okay.
18:57 So, let's jump into the the Q&A session, inshallah.
18:59 inshallah. Um,
19:06 let me open the presentation on my side quickly.
19:18 All right. So, explain the role of
19:21 captain in autonomous deployments and
19:24 how does it in integrates with the Argo rollouts.
19:37 I cannot hear you. Abdullah,
19:40 >> can you hear him now?
19:42 and lot of noise.
19:45 So, uh just dis just uh turn off the the
19:47 fan. We'll be done in 30 minutes. Don't
19:49 worry, you won't be sweating. Don't
19:51 worry. Okay.
19:54 >> My uh my fan is on pro. I'm not turning
19:55 on it.
19:58 >> Okay. So, keep the volume slow because
20:00 as you know sometimes there's a lot of
20:02 background noise and other things coming up.
20:03 up.
20:04 >> There's a lot of white noise, right?
20:07 Okay. So explain to me the role of
20:10 captain in autonomous deployments and
20:12 how does it integrates uh with argo rollouts.
20:14 rollouts.
20:16 Argo rollouts is used for traffic
20:18 shifting from uh traffic shifting uh
20:20 using deployment strategies like Henry
20:23 and blue green deployments while captain
20:25 is you uh while captain is uh
20:30 intelligently watching the uh watch the
20:34 matrix uh like SLOs's uh uh if the if
20:37 some uh if uh something if something
20:41 goes uh if error increase uh captain
20:44 will uh captain will trigger uh captain
20:48 will trigger and alert the uh uh it will
20:53 uh stop the uh uh roll out process.
20:59 >> So roll back. How does it uh
21:02 how does it executes you know uh like
21:06 for example captain it automates the it
21:08 automates what in the
21:09 >> uh in the
21:11 >> captain uh
21:13 uh like captain is intelligently looking
21:17 for a deployments uh based on SLOs's uh
21:19 we have set like error uh latency and
21:22 error rate uh it's uh
21:25 >> we we call those deployments as quality
21:27 gates and deployment,
21:29 >> right? We call them in the industry, we
21:31 call them quality gates and deployment decisions,
21:32 decisions,
21:34 >> right? And deployment decisions are made
21:37 using what?
21:38 >> Mention SLO. >> SLO
21:46 like error, latency and error rate.
21:49 >> SLO matrix. SLO matrix. These are all
21:51 matrix, right? It can be error matrix
21:52 and different. So, but those are known
21:54 as SLO matrix. Okay. Continue. >> Yeah.
21:56 >> Yeah.
21:58 >> So, that's what they do, right? >> Yeah.
21:59 >> Yeah.
22:02 >> Okay. Uh, what does captain evaluate?
22:03 I'm still not finished with the first question.
22:10 >> Like, uh,
22:12 captain uh, captain will uh, stop the
22:14 captain uh, if something goes wrong,
22:16 captain will stop the the roll out forces.
22:18 forces.
22:20 >> I think someone was just driving by.
22:22 >> Yeah. Sorry.
22:23 >> Okay. Sorry.
22:26 >> Uh like captain uh stop the roll uh stop
22:29 the deployment or roll it back or save deployment.
22:32 deployment. >> Okay.
22:38 So what is the purpose of the blackbox
22:40 exporter in this architecture that you
22:41 just showed? What does it do?
22:45 >> Yeah. Uh blackbox exporter uh is used to
22:47 uh is used to send uh send the fake
22:49 traffic to our services to check the
22:51 availability of uh to check the
22:55 availability and uh uh before uh before
22:58 it uh before the user notice and uh
23:07 Explain to me uh blue green architecture
23:11 deployments in the context of autonomous
23:14 deployments. What what is that?
23:17 >> Uh in blue green deployment uh traffic
23:19 shift all at once uh to the new uh to
23:22 the new deployment. Uh while in Kendry
23:26 we uh while in Kenry uh we don't uh uh
23:29 shift all the traffic all at once. Uh we
23:32 uh uh we make create a small amount of
23:36 users uh and then uh slowly
23:38 slowly shift traffic. If everything's
23:51 >> So, uh if I'm a retail merchant, right,
23:53 e-commerce merchant and
23:55 and
23:57 the network cannot be cannot go down
24:00 because we have uh lots of customers
24:02 buying in right now because of Black
24:06 Friday, right? Then we found out our
24:07 engineers found out that there's a
24:09 problem and they need to remediate the
24:12 problem. It can be a zero day
24:15 vulnerability. It can be something else,
24:17 misconfiguration item, whatever it is,
24:20 right? And they found out uh that there
24:23 is a fix for that. How would you use
24:25 blue green architecture to deploy the
24:40 how would you do that?
24:43 >> Uh so can you repeat?
24:52 for example
24:54 gap for example right and what is
24:56 happening is customers
25:04 and I cannot allow any downtime but the
25:06 engineers the site reliability engineers
25:21 What are they going to do? What are they
25:23 going to do specifically to redirect the
25:26 customers? Make sure everything is being
25:28 done while you're using Kubernetes or
25:31 using some of the other things.
25:36 >> Uh so we will use Argo rollout uh for uh
25:37 for our blue green deployment
25:40 strategies. Uh we will slow uh ship the
25:43 traffic from uh like uh if one of our
25:46 deployment have they fault uh like uh
25:49 like you said uh if uh we will shift uh
25:52 we will shift them from uh into a uh
25:52 previous deployment.
25:54 >> How do you shift? How do you shift? What
25:55 does shift mean?
25:58 >> Uh we will shift during blue green uh
26:00 via Argo rollout.
26:03 >> No, Argo comes later on. How do you shift?
26:05 shift?
26:08 You use uh load balancer. uh
26:09 >> to yeah
26:11 >> direct the traffic to the deployment
26:13 which is the green deployment away from
26:15 the blue blue deployment right the
26:18 customer betab
26:20 you are redirecting the new customers
26:23 over there right then
26:25 once those customers once and then what
26:27 do you do to move those customers on
26:28 those customers are done what do you
26:36 what do you call that process that
26:43 It's known as draining the traffic.
26:54 Okay. That's known as draining the
26:56 traffic. Okay. >> Okay.
26:57 >> Okay.
26:59 >> It's tough. Yeah. The topic is tough,
27:01 right? So be easy.
27:03 >> Uh differentiate between horizontal
27:05 ports autoscaler and vertical part
27:09 autoscaler. HPA and VPA what is it?
27:09 Explain to me.
27:13 >> Uh yeah, HPA is a uh it's a horizontal
27:15 port scaling uh based on CPU and memory
27:20 while VPA uh is uh uh scale based on uh
27:24 based on the resources uh of our board
27:26 uh like resources uh request for the
27:29 resources to adjust uh to adjust the board.
27:31 board. >> Okay.
27:42 uh
27:46 how do you combine them together?
27:48 >> Uh combine what?
27:51 >> The horizontal
27:53 pod autoscaler and vertical pod
27:54 autoscaler. Right? How do you combine
27:57 them? If I'm a if you're working as an S
28:01 sur engineer, how would you do that?
28:05 uh server will uh like uh implement both
28:10 uh like HP and BPA uh uh as uh one as a
28:13 CPU and uh and a memory uh and one for
28:15 uh for uh like uh
28:17 >> I understand I understand that partially
28:18 I understand that part
28:22 >> what I'm asking you is like in the S
28:35 Do you know uh ke
28:35 ke
28:36 >> to scale? >> Yeah.
28:37 >> Yeah.
28:38 >> Yes. What is
28:42 >> Kea is beyond about that. Uh it is like
28:45 uh uh scaling based on business events.
28:48 >> Yes. like like the uh
28:51 >> so in S sur both are combined with kea
28:53 for eventdriven
28:55 scaling to handle unpredictable
28:56 workloads I'm looking at your
28:58 presentation that's how you do it right
29:00 because horizontal autoscaling and vert
29:02 horizontal autoscaling is all is all
29:04 about what CPU and memory
29:05 >> CPU and memory
29:07 >> vertical autoscaling is about resources
29:10 requests limits for existing pods right
29:12 because it is you know this is what we
29:14 have right now in our side as well every
29:18 Say we have reserve instances which are
29:21 running in AWS
29:24 and when the load increases
29:27 it goes from a pod of five to a part of
29:31 6 10 12 15 it keeps on going right
29:35 >> and that is designed by using ka
29:38 kubernetes eventdriven autoscaling okay
29:40 remember that this is very important
29:44 okay all right so why would an s sur
29:48 Choose spot ocean or carpenter for cost
29:50 management. What are the benefits?
29:53 >> Uh sir, it let us uh choose the cheapest
29:58 instance available uh uh like uh spot uh
30:03 like uh it tells us the uh that uh like
30:06 uh the
30:08 instance available in the cheapest
30:11 price. uh we used uh it help us.
30:13 >> It optimizes the infrastructure cost by
30:15 leveraging spot instances. Whatever is
30:17 cheaper, we use that first, right? >> Yeah.
30:17 >> Yeah.
30:19 >> But it's not it's I I understand you're
30:22 sweating because of the fan. It's okay.
30:24 >> You'll be done. The scores are good.
30:26 Mashallah. Scores are good. Okay.
30:28 >> Words are coming. Good. Mashallah.
30:30 >> Don't worry. Don't worry. We're almost
30:33 there. Almost there.
30:36 >> Uh okay. uh what advantages does
30:40 distributed tracing with Jagger provide
30:42 over traditional logging?
30:45 >> So distribute tracing uh we use
30:47 distribute uh distribute tracing uh to
30:51 check uh however uh how request travel
30:54 from uh multiple uh multiple services.
30:57 Uh we used uh uh open telemetry and Jagger
30:58 Jagger
31:03 uh to uh like open telemetry uh uh tele
31:05 uh use the telemetries of matrix logs
31:08 and matrix logs and Jagger uh displays
31:10 them. Uh Jagger is the back uh back end
31:13 and the UI that visualize them uh that
31:14 visualize them.
31:16 >> Good mashallah. You know like Jagger
31:19 traces the entire life cycle of the
31:21 request across microservices. That's
31:23 what it does. and uh you know revealing
31:25 what revealing performance
31:27 >> it help us debug the
31:30 >> latency issues all of these things good >> uh
31:32 >> uh
31:33 the thing is like this is very important
31:35 for you to understand you have to do the
31:38 alazra labs because the labs is where
31:41 you will be doing these things labs
31:42 cover all of these things and their labs
31:45 are tough right so that will prepare you
31:47 for the physical interview when you have
31:49 to go in today I was reading somewhere
31:52 that organizations have started calling
31:53 people to come on site for the interviews.
31:54 interviews.
31:57 >> Yes. come on side because you know when
31:59 when you're on site you will be put on
32:02 the spot there to explain you know how
32:04 does the autonomous feedback loop
32:07 architecture in S sur works so many
32:09 things you know okay explain to me what
32:11 is self-healing
32:16 >> selfhealing uh yeah if something like uh
32:18 uh something breaks or something uh
32:21 happened our uh we have the tools like
32:24 pad duty that uh trigger uh that also
32:26 inform home the team and also trigger
32:30 the play uh uh like uh tools uh like NCL
32:33 AWX that fix the issues automatically.
32:35 Uh that has the job to uh restart the
32:39 board uh clear caching and fix the
32:50 You can tell everyone you know I
32:52 literally sweated throughout the whole interview.
32:54 interview.
33:01 >> You tell everyone it was such a tough
33:03 that I had to do everything, man. Like
33:20 Okay. Um
33:22 how would you implement predictive
33:24 scaling for black Friday traffic using
33:28 ka and prometheus metrics?
33:30 Uh we uh we use predictive scaling uh
33:35 that is uh based on uh our past data uh
33:38 that is based on our past data to uh
33:40 increase the resources on the uh the
33:46 future. uh like uh it's an uh like uh if
33:49 uh uh a traffic uh increase in every uh
33:53 Friday night uh uh every Friday night uh
33:55 uh it will uh it will increase in the
34:12 So wouldn't you be looking at historical
34:15 data traffic and ML models for for
34:17 demand forecasting
34:19 when you're doing the predictive you
34:24 know scaling for Black Friday
34:28 >> uh like we use the our past data uh
34:28 right like
34:30 >> this is very important you know yes we
34:33 use ka can trigger all of those things but
34:33 but
34:35 >> based on what
34:36 >> you know today
34:38 >> based on the historical data you know
34:41 al-Basir is is our data lake right every
34:45 system log goes there
34:47 so we using different visualization
34:49 tools to look into what is happening and
34:51 we were able to find out that where the
34:54 people are having the most common
34:57 issues accessing the containers on the
35:00 cloud and they were all clumped into
35:02 specific area so we were able to see
35:06 okay this is the only place you know uh
35:08 in Pakistan, in India, in Bangladesh
35:10 where the people are having problems.
35:13 They're not having problems in USA, UAE,
35:16 Saudi Arabia, Australia, New Zealand,
35:19 Germany, France. So, we were able to
35:20 pinpoint it, right? >> Yeah.
35:21 >> Yeah.
35:23 >> And this was all because of his. So,
35:25 this is our historical data. Now, if any
35:28 customer comes and say, I have a
35:29 problem, I have this, I have that. So,
35:31 we can say
35:33 there are 10 customers having the same
35:35 problem as you. are you using the same
35:38 internet provider and if he says yes
35:40 that's your problem
35:42 >> right so historical data is very
35:43 important that's how we are
35:46 troubleshooting our own issues as well
35:47 you know when you go into the
35:49 organization you have to dig dive you
35:51 know dig take a deeper dive into the
35:55 data to see what is happening okay and
35:57 that's how you do it
36:00 >> describe a CI/CD integrated performance
36:03 testing setup with Ksix or an e-commerce platform
36:05 platform
36:08 uh we uh we use KS6 for our uh
36:10 performance testing uh during CI/CD
36:13 pipelines. Uh uh before uh before
36:16 deploying a new version uh uh we use
36:20 Ksix to simulate user uh user traffic uh
36:24 that will uh that uh uh that will let us
36:27 now uh that will uh that will check the
36:30 deployment uh before uh that it can uh
36:46 So why is chaos engineering essential in
36:54 >> Uh why is chaos engineering
36:56 >> essential? Why is it important in autonomous?
36:56 autonomous?
36:58 >> Yeah. uh
37:02 uh in uh like in traditional S sur uh we
37:04 wait for some uh something to break down
37:06 uh and then fix it manually. But in
37:09 autonomous uh we use chaos monkey that
37:11 uh uh randomly break down the parts of
37:14 our system uh uh it sounds risky but it
37:16 is very helpful. So we can see if one of
37:19 our service uh goes down other also
37:22 others will keep working. uh and will it
37:24 will also let us uh it will also confirm
37:27 that our autoscaler and uh encable
37:41 Okay,
37:43 the exam is over. You can turn on the
37:44 fan if you want or you can just go
37:46 through it if you like.
37:49 >> Yeah, I'm sorry.
37:51 >> Turn on the fan. It's getting hot
37:52 >> now. I'm good now.
37:53 >> Okay. Good. Okay. >> Yeah.
37:54 >> Yeah.
37:56 >> See, after some time a person get
38:15 >> Okay. Uh the interview was good.
38:20 Mashallah. uh you secured 22 out of 25
38:22 in your presentation. That's 88%
38:24 mashallah. Okay. >> Okay.
38:25 >> Okay.
38:28 >> Uh in the uh
38:32 in the exam uh the Q&A uh that was the
38:34 presentation. In the Q&A, you secured
38:37 23.5 out of 25, which was good.
38:41 Mashallah. That was 94%.
38:47 So the overall score is 88 + 94
38:50 is coming out to be 91% mash. >> Alhamdulillah.
38:51 >> Alhamdulillah.
38:54 >> Alhamdulillah. Okay. So, so this is my
38:57 recommendation to you. It's very
39:00 important. Please pay attention and make
39:03 sure that you listen. Uh I'm going to
39:05 quickly just give me one minute. Update
39:12 So just need to insert these numbers in
39:14 there because then they go finalize and
39:17 then system will take it over. Okay. So
39:20 here's what needs to be done. Okay.
39:24 Okay. This is my recommendation for you.
39:26 So pay attention to what needs to be
39:28 done so that you can do all of those
39:30 steps and you can make sure that those
39:33 things are carried forward. Okay. If you
39:36 log on to the portal,
39:40 you will see multiple things. Okay. >> Yeah.
39:41 >> Yeah.
39:43 >> I'm going to show you that. I'm going to
39:47 show show it to you. So, first of all,
39:48 uh you have already secured the job, right?
39:50 right?
39:51 >> Yes, sir.
39:53 >> Mashallah. Okay.
39:55 >> When did you started the preparation for
39:58 the exam?
40:00 >> Uh like uh this presentation.
40:03 >> Are those your kids
40:07 or are those in the neighborhood?
40:09 Uh I don't understand sir
40:11 >> people who are you know I'm I'm hearing
40:13 kids shouting in the background. Yeah,
40:17 it's uh it's outside neighborh.
40:20 >> So mah you secured 91%. So go outside
40:23 and tell them that happy these are the
40:27 sweets for you. Okay. So give them
40:28 >> don't scold them. Don't do don't say
40:31 anything. This is the sunnah of
40:33 >> used to be happy. >> Okay.
40:34 >> Okay.
40:37 >> Uh all right. So
40:39 don't forget to do that. You have to do
40:40 that inshallah. Okay. So give them
40:42 something like that. and they will be
40:46 very happy in all right so zak uh so
40:48 here's the thing that I have for you right
40:50 right
40:52 uh when did you started your studies
40:54 with us
40:59 >> uh I'm started studying uh from uh 2023
41:02 >> so 23 which month
41:07 >> I think it's May uh or June I don't uh
41:09 really remember
41:11 >> so mashallah you moved there you started
41:13 your preparation, you started your
41:15 studies. Prior to joining us, did you
41:17 know anything about all of these things
41:21 that we discussing right now AI ops,
41:23 devops, sysops, any of those things?
41:28 >> Uh no sir. Uh I just have only uh the
41:30 knowledge about the cyber security. uh
41:32 because at the first uh from the
41:35 starting uh when you start uh teaching
41:40 uh in Pakistan or at uh 250 uh 250 250
41:44 rupees uh like uh per year
41:45 >> there was no accreditation there was nothing
41:46 nothing
41:47 >> there was nothing
41:49 >> there was no cloud there was nothing >> yes
41:50 >> yes
41:52 >> that's how we started that's how that's
41:54 a long time ago right >> yes
41:54 >> yes
41:56 >> you started uh learning because I'm
42:00 looking at your resume right and you did
42:02 your intermediate in 2022.
42:05 You did your metriculation O level and A
42:07 level equivalency and then you went
42:11 >> 2023 and you completed that and almost
42:13 one one and a half year one year and
42:15 five months, right?
42:16 >> Yes sir.
42:19 >> Right. Uh something like that. Oh almost
42:23 like you studied AIS for how how long?
42:25 Uh sir first of uh first I start
42:28 studying uh level three cyber security
42:30 then I will move to
42:34 uh devops uh level five uh and when I
42:38 see the uh p uh fee structure is not uh
42:42 that much high so then I move to AIOPS.
42:44 >> Okay. So now mashallah you have
42:45 completed this thing and you have
42:48 already secured a job right? Uh
42:49 >> yeah right.
42:51 >> Uh this is somehat government
42:53 organization right now. Uh yeah,
42:55 government organization NAV.
42:56 >> You'll be doing you'll be doing what
42:58 over there like you'll be teaching
43:01 >> uh teaching docker certified associate. >> Okay.
43:02 >> Okay. >> Teacher.
43:03 >> Teacher.
43:05 >> Okay. So they're paying you a six figure
43:06 salary. How much?
43:10 >> Uh yes. Six figure salary. 150.
43:12 >> Mashallah. That's good. That's good. So
43:15 whatever you invested here, you'll be
43:16 able to take it out inshallah.
43:18 >> Yes sir. >> Yeah.
43:18 >> Yeah.
43:20 >> Inshallah. Okay. So >> sure
43:20 >> sure
43:24 >> this is my recommendation to you. Okay.
43:26 If you log on to the portal, if you log
43:27 on to the portal, you will see
43:30 something. Uh this is very important for
43:32 you to understand. When you log onto the
43:34 portal, you will see alazar,
43:37 right? Yeah. So this part one, my portal
43:38 is different because I see everything,
43:41 right? So you will see alazak part one.
43:43 This is important. And then you will see
43:45 alazak part two. Alazak part one is part
43:47 of the course. So you need to complete
43:48 this. You need to complete all of these
43:50 labs and these labs will keep on coming.
43:54 So you complete all of them. But then
43:57 this is alazak part two. Very very
43:59 important for this for getting the
44:01 diploma. You need to complete this these
44:04 labs, right? You need to complete all of
44:06 them. This is very important for you.
44:08 >> Uh and they will keep on coming. You
44:10 will see that they will keep on coming
44:16 etc etc. But like to get the you know uh
44:18 be ready for the job be ready for the
44:21 teaching all of those things your whole
44:24 job is here containerization do
44:26 >> but you can you will be able to teach
44:30 your students the whole QP astronaut
44:33 right with all
44:36 >> I can't because uh because the
44:39 government uh choose uh what I need to
44:42 uh what they want to teach their uh students.
44:43 students.
44:46 >> Okay. So there's a specific protocol
44:48 >> but whatever it is like I'm not sure like
44:49 like
44:50 >> if you guys will be able to provide them
44:52 with the labs and everything and the
44:54 containers right
44:56 >> uh just like we do if you click here on
44:58 dockers right this is your whole course
45:01 right uh but the good thing is there are
45:02 labs here
45:05 >> the labs cover everything right it
45:07 covers all the aspects so if you click
45:10 on here you see lab three it starts with
45:13 lab 1 2 3 4 5 6 and the labs keeps on
45:17 going until you have almost
45:21 you know 100 labs more than 100 labs 110 labs
45:23 labs
45:26 running Apache CFKA with Docker Compost
45:28 running Docker containers with Microsoft
45:30 Jour like it keeps on coming and coming
45:33 and coming right and this is what
45:34 prepares you
45:38 >> you know for what you have to do right
45:41 this is just one example so you can see
45:43 the lab is there what do you have to do
45:45 in the lab. The container is already
45:47 ready. You start the lab, you do the
45:50 lab, the AI assistance is already ready.
45:53 Move here. Right? So yes,
45:54 >> everything is there
45:57 >> and this labs contains check docker
45:59 version, check docker compose, check all
46:01 of those things. Make sure download and
46:04 configure clear. It goes through all the
46:07 details. These are beginner's lab. These
46:09 are not even intermediate and advanced
46:11 which we are about to release. Right? So
46:14 in Allah zak you will get most of these
46:16 labs. The other thing that you will get
46:20 is basically you know part two which
46:22 covers many other aspects as well that
46:25 you'll be covering. Right? So if I go
46:29 back and uh quickly
46:34 go here right uh let me see give me one
46:37 minute. Hold on.
46:39 Just want to make sure I'm on the right screen.
46:46 close.
46:50 Yes. Okay. So, this is the one. So, if
46:52 you go back on the course, you will see
46:54 something very important. You will see
46:58 here progress and you will see uh courses
47:01 courses
47:05 very very important for you to focus
47:09 right Linux open source it covers
47:12 hands-on labs as per the job description
47:14 not for passing the exam you can pass
47:16 the exam by studying them as well but
47:18 they are based on job description
47:20 >> see red hat the whole redhead repository is
47:29 whole altogether more than 45 plus exams
47:31 open shift and everything and then open
47:32 shift and all of those things were there
47:34 on top as well. Yes,
47:39 >> this is what prepares you for the market
47:41 because whatever course you you know for
47:43 example anible automation if you click
47:45 on it you will find the course but you
47:48 will also find all the labs all the labs
47:51 and there are handful of labs right and
47:52 all of these things there are lots of
47:57 labs you can pick up a lab lab 11 lab 12
48:00 they're like more than 20 labs right 20
48:03 plus labs something like that
48:04 >> you can
48:06 This is very important for you. If you
48:08 go through all of these things and if
48:12 you focus your time and effort, you'll
48:15 be able to pass any interview whatsoever
48:26 What do you have to do next? You have
48:30 scored 91%.
48:32 Laz program complete complete your Lazak
48:35 program become QB astronaut
48:37 that teaching job if the government is
48:41 good is okay but this is not your future
48:42 >> yes sir
48:43 >> will get the money back that you have
48:45 invested you did your inter intermediate
48:47 then you did this and you have completed
48:49 your bachelors alhamdulillah you can go
48:53 forward but think big you know think big
48:55 right everyone
48:57 >> so there are seven levels of ja right
49:01 >> everyone wants to Which level?
49:02 >> Well, at the level seven.
49:05 >> Why? Why is it is there a difference
49:08 between level one and level seven?
49:11 >> Yes sir. The big big difference.
49:14 >> Big difference. Right. Level seven
49:16 >> is first of all just beneath alas the
49:20 jah ofam and above the ja of rasoolam is
49:23 the throne of Allah. Right? So Allah is
49:25 watching all of us. But what you get in
49:28 level one you don't you know level two
49:30 you get more level three you get more
49:32 level four level five level six and
49:35 level seven level seven is the ultimate
49:37 it can't get better than this it that's
49:40 the best right so everyone wants to
49:43 >> if you have a option to fly there's
49:45 economy and then there is business and
49:46 there's first class everyone wants to go
49:50 into the first place why
49:53 >> because it's better
49:55 >> it's a You have but you have to pay
49:57 more. You have to work hard to get earn
49:59 that money to be able to do that, right? >> Yes.
49:59 >> Yes. >> Similarly,
50:00 >> Similarly, >> right?
50:00 >> right?
50:04 >> Yes, you're going in business right now
50:06 from the economy to the business because
50:08 you secured a job but you need to go to
50:10 the first class you need to go to level
50:11 seven. For that you have to work hard.
50:20 but don't stop your learning. Learning stop
50:23 stop
50:25 you stop your learning. You will be left behind.
50:26 behind.
50:28 You have to focus on
50:30 >> you have to focus on the most important
50:34 thing is the core integration framework.
50:41 I'm going to pick up quantum computing.
50:42 I'm going to pick pick up energy
50:44 systems. I'm going to pick up like 5G,
50:47 6G communications. I'm going to pick up
50:50 this. Any one of them you can pick up.
50:52 But do it.
50:54 Practice, practice, practice. All the
50:56 labs are there. You don't have to study
50:58 these courses anymore. You can go on
51:00 cloud and you can buy the labs from
51:03 there. Just the labs and the videos for
51:07 the specialization. But go and do it in
51:09 >> Okay. Okay. And also target to go
51:12 outside once
51:16 gather some money and then move outside
51:19 or a master's degree
51:31 you will you've already secured one job in
51:32 in
51:36 >> but inshallah what I'm saying is in this country
51:42 you can't go anywhere further. >> Yes,
51:43 >> Yes,
51:58 say
52:00 91% may meet up all the other
52:04 requirements you can already uh secure a
52:07 bank financial statement from us
52:09 requirements complete you already did
52:13 that was a part so you already did that
52:20 meet all the other requirements and then
52:23 you can go in
52:25 >> just gather some of the fee if but
52:28 before You go get all the skills right. >> Okay.
52:30 >> Okay. >> Okay.
52:30 >> Okay. >> Okay.
52:31 >> Okay.
52:33 >> If ask your parents to make special dua
52:42 >> to make me inshallah
52:44 for your time. I'm really happy for you.
52:46 You did very well. Mashallah. Okay. uh
52:48 it will take four to six weeks maybe 8
52:51 weeks to submit everything get the
52:52 certificate and everything but we can
52:54 give you a provisional letter in the
52:55 meantime in okay >> sure
52:56 >> sure
52:57 >> one document sir
53:00 >> which is very important for you to know
53:02 and have which is this I'm going to send
53:05 you this link here so that you have it
53:09 as well uh on the chat this is
53:11 qualification can cross boundaries right
53:14 so you did your diploma at level
53:16 >> this is level six diploma
53:18 which is equivalent to bachelor's degree
53:20 with honor.
53:22 >> So this is RQF level six. This is what
53:23 we for
53:26 and this is what you did right after this
53:27 this
53:28 >> there is level seven you can take
53:30 admission in level seven many students
53:32 from uh different places are leaving
53:34 outside right so >> yes
53:35 >> yes
53:36 >> I'm telling you inshallah you have a
53:38 good opportunity you have everything
53:40 secured work hard may Allah give you
53:42 more success do you have any questions
53:43 for me
53:46 >> uh sir can I uh
53:49 uh like do CISSP certification
53:52 >> no need no need I
53:57 like like 725
54:00 do the specialization
54:02 go and go pass QB astronaut that's much better
54:04 better >> okay
54:06 >> okay qualification
54:11 when you secure a job let the employer
54:20 no visa the companies will look at your GitHub
54:21 GitHub
54:23 Do the labs, put them on the GitHub. I
54:25 did this. I did this. I did this. I did
54:27 this. I did this. These are my projects.
54:30 Do all of those things. Start blogging
54:33 on LinkedIn.
54:33 >> Yes, sir.
54:36 >> That's much better.
54:49 That's much better. Inshallah. Become QB
54:50 astronaut. become golden cubin at all
54:53 the labs you will find all the labs
54:56 everything is ready for you to consume
54:57 it's like Jenna you know like you don't
54:59 have to do any installation you don't
55:00 have to cook you don't have to do
55:03 anything and the food is there for you
55:05 consume it and we're telling you how to
55:07 eat it as well what commands you have to
55:09 run what do you have to do everything is
55:12 there why are you looking for hardship
55:15 passing CSSP is not worth it
55:18 >> sir uh for getting uh QBS to node uh uh
55:21 did I need to enroll again. Uh
55:23 >> as long as your membership is active,
55:24 you can
55:24 >> Yes, sir.
55:27 >> get the benefit from Al Razak Labs.
55:28 >> Sure. Okay,
Chrome Extension