How to Scale Your Business in the Cloud

0:05 i guess we are live yeah hello and welcome everyone to our 0:10 webinar my name is salman iqbal and i’m joined with tennis and today we are going to be 0:16 talking about business scalability uh and uh yes welcome everyone welcome 0:22 tennis tennis how are you doing this morning very good very good um my name is tennis yeah that’s my real 0:27 name and uh and simon how long have you been in the business now uh beginning of 0:33 this year so uh nine months nine months okay well this is my 42nd year in the business 0:40 oh okay when you talk about business oh yeah so i’ve been with up here for nine months but i have been in the business 0:46 for about eight years okay okay so we’re we’ve both been around for a while um 0:53 so anyway i uh simon and i worked together and so 0:58 this is the uh the the context that we’re uh pursuing here is just two professionals 1:05 talking and uh that’s really the the the uh arrangement we have today 1:12 yeah perfect yeah all we’re going to be talking about scalability in in your business and how it relates 1:17 to technology and while we’re talking please feel free to click around and submit your questions you’ll see a panel 1:23 on your right i believe where you can submit your questions anytime you have any questions let us know what we 1:30 discussed today we have a blog to go with it that touches upon the discussions that we’re going to have so 1:36 you know we’ll we’ll be sharing that with you and yeah also please submit any of the feedback anything else you’d like 1:42 to hear about after this uh any other topics that you want we can we can you know we can pull it in on our on our 1:48 expertise it’s not just it’s not just tennis and myself we have a whole bunch of people who help us do this so well 1:54 maybe we should do our intros uh tennis would you like to go first and introduce yourself give us a bit more about you 1:59 who you are sure tell us about your swords behind you everything oh yes this is the i do martial arts and those are 2:06 swords those are real swords um i uh i’ve been in the business as i said 42 2:12 years this is my 42nd year done everything as far as working in it i run cables i’ve installed things i’ve 2:19 traveled to to customer sites teach classes i’ve been a developer and lately was a devops guy so i’ve 2:27 been there and done that i can honestly say for most most areas of the i t world 2:32 excellent excellent thank you thank you very much tennyson yeah my name is salmanik baal i don’t do any martial 2:38 arts uh unfortunately so i would not be messing with tennis anytime soon yeah i am based in the uk 2:44 in in wales uh i have been working in it for about eight nine years so i’ve worked in aerospace finance banking 2:51 government you name it and i i started my life as a developer then i got into devops cloud 2:58 devops and currently i work as an amalox engineer it’s all about taking your machine learning and data science 3:04 workloads and running them in the cloud efficiently and uh at scale so that’s what i do scale is the word of the day 3:11 yeah isn’t that right tennis yes sir and i’ll with that segway into talking about 3:16 one thing that we’re going to make real clear about scalability is is that there there isn’t any question and among them 3:23 we’re talking about scalability and the cloud there isn’t any question that’s that 3:28 technical scalability traffic scalability is achievable and well understood and there is no issue with 3:36 all the the the providers being able to handle your traffic volumes that’s not the scalability we’re here to talk about 3:43 we’re here to talk about a different tack on that same subject which is 3:49 scaling your organization which is something that people don’t talk about an awful lot is now that i’ve got this 3:56 wonderful environment on for example aws and pretty soon everybody wants to use 4:02 it how do i scale my organization to handle it and that’s really the context of what we’re going to be talking about 4:07 today yeah perfect so we work for appvia and apia work with lots of organizations 4:14 we help other organizations scale up their business from technology point of view and from business point of view so 4:21 you can check us out in april.i o we’ve got lots of solutions we’ve got lots of bright people 4:27 and what we do is we help other organizations adapt cloud native technologies and accelerate their software development 4:34 and delivery and that’s the point right we want to get the product in front of your users yeah how do you do that is by 4:41 accelerating your development and scaling your business and i guess that is exactly what we’re going to talk 4:47 about today yeah and then yeah god’s sorry tell us a bit more tell us a bit more The classic scenario 4:53 about this well the the classic scenario is hopefully the scenario is 4:58 that you you make a sojourn into the cloud and you make uh you bring up for example 5:05 kubernetes and you use one of the pre pre-configured 5:10 kubernetes offerings from the cloud providers like aks for azure or eks for 5:17 aws and you bring it up and it works fine and you find that everything is as 5:23 advertised um and then you’ve you know continue on 5:28 and a couple months go by and you’re a little more successful and more people are interested and somebody says well can i have a cluster too and they say 5:35 yeah sure and before you know it you’ve you’ve got to uh uh 5:42 maintain lots of different clusters even within the context of these eks aks 5:47 offerings that’s a great deal of effort um and as salman will tell you that’s 5:53 that’s not something that’s easily done in terms of the number of people that are going to be involved 5:59 so what happens is you over time find yourself at least tempted to throw 6:05 bodies at this problem so that you have more and more people that are going to make be maintaining it does that make 6:11 sense salmon am i yeah so yeah so i think what you’re saying if i if i understood this correctly 6:18 we we are an organization and we are in this process of acceleration and we need to we’re developing our products and 6:24 we’ve got all these users let’s say let’s call them developers or support people or whatever they are and 6:30 they need to use the resources that could be kubernetes clusters could be virtual machines it could be anything 6:36 right it could be anything and they come in they sign up and they say all right i want i want a cluster let’s let’s go with kubernetes cluster 6:42 because that’s the hot thing right now right yeah i want to sign up i want a cluster i guess there is a ticket somebody creates them a cluster somebody 6:49 else joins the company you start creating clusters so basically we’ve got tons of tickets now people 6:55 trying to create clusters for as and when people get onboarded it’s not just i’m guessing it’s not just Creating clusters 7:00 creating clusters right it’s like what permissions do you have exactly kind of rules is that is that right exactly yeah 7:07 what you find uh is that you what people most of the time do is they retrofit existing ticketing 7:14 processes into this new uh arena and then they say well just submit a 7:19 ticket and we’ll fire up a new cluster for you and what you find over time is that 7:26 it gets harder and harder to keep up with those tickets because again we’re assuming you’re being successful and if you are successful 7:32 you are going to have more need for these facilities on the cloud okay and that success 7:39 breeds a real interesting problem where you simply can’t keep up with the 7:44 requirements excuse me requests that keep coming in um and that’s where things tend to go off 7:51 the rails because the the the temptation is that’s okay i’ll just hire another 7:57 administrator because after all all i’m doing is i’m bringing up a cluster by hand 8:02 okay um so that’s that’s the temptation that’s what not to do okay so basically 8:07 what if i were to summarize the discussions we’ve had uh we’re an organization and we can 8:14 handle all the technical stuff as in we can handle the traffic that’s coming onto our websites and you know the cloud 8:20 provider as you’re saying has provided us the answer we’ve got load balancers we’ve got clusters that they’re running in people who’ve got the right 8:26 permissions but one aspect that we don’t think about is scalability of the organization that’s right and how can Scalability of the organization 8:32 you handle the success in the cloud is that that’s what he’s saying right that’s right it you can be a victim of your own success and the the the 8:39 temptation it says that’s okay i will just hire what will happen is is this is that you will get complaints you know my 8:45 people need a cluster and you told us last week you’d have it and it still isn’t there what’s the problem that’s okay i’ll get another couple of 8:51 administrators or i’ll retrain somebody who’s already in the organization to be an administrator because all you have to 8:57 do is stand all you have to do quote unquote is stand up a cluster and you find that that is a much bigger effort 9:04 as time goes on because you will have particular needs that uh people well i 9:09 need these permissions and i need a cluster that looks like this and those kinds of things yeah and no i’m sorry 9:14 good no no that’s correct because like when we say give me a cluster you need to create everything that goes around it 9:20 right all the networking all the security all the resources of like where is where are my images gonna get stored what are 9:26 the permissions what applications can talk for what application that’s right and that you’re saying doing it manually Manual processes 9:32 is a painful process and i’ve done that in the past and a hundred percent know every time i do it i can never do it the 9:38 same i did it the first time around that’s that’s right and then that’s an excellent point uh even if you mandate 9:44 we’re going to build our clusters with these guidelines always and forever you know 9:50 our namespaces are going to be named this and the way they’re going to work is that and what will tend to happen is the 9:55 proliferation of manual uh guidelines and cheat sheets 10:01 that will outline how you’re supposed to do this and okay salmon i don’t know about you but i’ve never been able to 10:07 create two separate clusters that are identical truly identical 10:13 um and and and it’s not the failing of the technology so much is just the complexity of the 10:19 technology that we’re talking about um so what happens is is that that 10:24 again you you start to add people to this process and then you add guidelines for those 10:30 people to follow and then then mistakes start happening you know you told me last week you were 10:35 going to bring up a new cluster for me you did but how come the name space is named foo you know you know those kinds 10:41 of things um and that’s that’s doesn’t it doesn’t everybody have a food namespace 10:47 exactly you know it works for me development right you know the classic developers line 10:53 well it’ll work for me on my channel absolutely so so what what i’m here what i’m hearing is saying is 10:59 like now how not to solve a problem is not to keep doing what perhaps you were doing 11:04 before as in what was working before maybe manually might not work anymore and you’re saying The antipattern 11:10 if you keep throwing bodies at this problem you’re spending more money but you’re not getting where you wanted to 11:16 get to right exactly and ultimately it’s a losing proposition you you’re trying to surf on this giant wave of 11:24 requirements that people are and and that is an extremely dangerous 11:30 place to be because ultimately you can’t keep away you can’t keep ahead of it uh so um um 11:36 the the the now we’ve talked about this is the anti-pattern the the pattern should be 11:43 to go and try to find ways to automate this effort and have your people spend time on the 11:49 automation by the way one thing i haven’t talked about too is that if you hire a bunch of 11:54 intelligent highly trained individuals to sit around and do repetitious things 12:00 day in and day out they won’t stay very long yeah um that’s another problem that you 12:06 know you you are guaranteeing high turnover um uh with yeah and and people say that’s 12:13 just bringing up a cluster well just bringing up a cluster is a pretty complicated thing so if you if you turn it into a 12:20 repetitious uh effort then nobody’s going to stay and then people say well why don’t you 12:25 just do it with something like terraform which you can do um but we’ll get into that in just a 12:32 moment i’m sorry i’m getting ahead of myself getting ahead of that you know it’s an exciting talk you know it’s an exciting discussion so well 12:38 welcome everyone people who’ve joined us uh while we were talking welcome to the webinar we are talking about business 12:43 scalability once you have started using your technology the way you wanted to in the cloud you come at i’m gonna call it 12:49 crossroads right is that 12:56 and as tennis was explaining technically everything’s working fine but the problem is you’ve got all these users as 13:02 in your developers who are building these services you need to onboard them onto your system i need to provide them 13:08 the resources they need and if you were doing this manually beforehand it won’t work anymore is that a good 13:14 summary there that’s a good summary yeah and if you have any questions please feel free to pop your questions into the 13:20 chat and we will answer them uh we’ve got so tennis so far has talked about 13:25 how not to do i don’t solve this problem right yeah so all i hear is 13:30 uh i i give you some solutions tennis and i’m sure people want to hear some solutions like what what do we do in 13:35 this case we shouldn’t throw bodies at it because we’ll waste money yeah and people will 13:41 leave as you say and i completely agree with that um so what what should we do tennis well Level 1 Automation 13:46 the the in one word it’s automation uh but even then you have to be careful um 13:52 the the the initial um inclination would be say well i’ll just 13:58 write a bunch of terraform scripts nothing wrong with that it’ll work just fine 14:03 problem is you now have to maintain those terraform scripts 14:08 and and that’s something that people don’t realize that you you’re getting into is is not not every engineer is 14:16 going to be a developer but you’re requiring them to be a developer in order to to build and maintain telephone 14:22 terraform scripts okay again nothing necessarily wrong with that it’s just that you 14:27 have you have basically a priesthood that looks after kubernetes and then you 14:33 have a new priesthood that looks after the kubernetes scripts okay um and 14:38 again if that’s what you’re signing up for and you realize it that’s fine but the the long-term implication is is that 14:45 you will have these terraform scripts and the scripts are going to tend to vary depending on the on the on the 14:51 provider so you’ll have aws scripts and you’ll have azure scripts you’ll have gcp scripts um and you will have probably multiple 14:59 variations on each so before you know it you have a large library of scripts that 15:04 need to be one hopes documented and maintained yeah and the organization needs to 15:11 realize that is a long-term commitment yeah um so that’s that’s what i would call the sort of 15:17 um uh sort of level one way to do it is to let’s just write some terraform 15:23 yeah just write some scripts so yeah just in case people who are listening in if you’ve never not heard of terraform 15:29 before oh yes it’s an infrastructure as code uh way of doing things so as in rather than 15:35 me and you going into a cloud provider and clicking around and creating stuff we can write that in code and the point 15:43 about using in code is it’s repeatable the same script i run it can deploy things correctly and 15:51 yeah and basically what we’re gonna do is instead of you can write your own scripts but if you write your own 15:56 scripts you have to maintain them is that right that’s right and and that is something Level 2 Maintenance 16:02 that you’re you know everybody everybody’s thrilled on the day one process of getting those scripts 16:08 off the ground and getting them running that’s really wonderful the day two problem is what people don’t think about 16:15 which is maintenance and you know version changes and those kinds of things and 16:21 that’s that’s a long-term commitment as i said so you have to be absolutely aware of that now that’s like i said 16:27 that’s the sort of layer one by the way thank you for telling the world about terraform i just sort of assumed everybody you know uh 16:34 but thank you for uh for for level setting on that uh there are other 16:39 tools too but let’s terraform is the one that’s generic and works across all those platform providers that we talked 16:44 about um the the next step another approach 16:50 in the automation uh umbrella is to buy a tool or buy capabilities and 16:59 that’s where i will to our own horn which is to say that’s what appvia offers is the ability to off the shelf 17:07 uh offer these kinds capabilities and you do not have to be a developer in order to run these things 17:13 and you have and by the way unlike unlike terraform or 17:19 other scripting tools the support process isn’t really there 17:27 for them you have to get on different message boards and ask questions if something blows up and they’ll say well 17:33 that’s too bad why don’t you try this there’s no formalized support process 17:38 unless you pay a lot of money for terraform um 17:44 on the other hand if you have a pre-packaged arrangement like you have with the appvia tools 17:50 then you you actually have somebody to call and say i have a problem could you help me with that uh and enterprises 17:56 particularly like that nobody likes the idea if you’re sitting by and some vp sitting behind death says how come such 18:03 and such broke last night and we couldn’t bring up terraform and the guy goes i put a message on the message board and i’m hoping to get an answer 18:09 any minute uh that doesn’t go too far uh oh 18:14 i’ve lost jeff uh uh salmon don’t can’t hear you 18:22 can’t hear you dude 18:29 can’t hear you 18:34 no don’t hear nothing i’m Technical Glitch 18:40 oh that sounds better can you hear me now yeah all right 18:45 there’s there’s always at least one technical glitch in every webinar but 18:51 you you can hear me now right i’m i’m i can hear you just fine all right i can hear you just fine so basically what we’re saying so far what i was saying 18:58 why you couldn’t hear me like i’m sure people can if if you can’t hear us please let us know um but so 19:04 what what i was gonna summarize what you were saying is if you use any of these tools you can 19:10 automate stuff of course you can use terraform you can use pollumi uh anything you like if you use any of these tools first you need to have the 19:17 knowledge of the cloud provider that you’re using yes sir yes gcp whichever you’re using and uh 19:24 then you have to look after the scripts that you wrote yes that’s quite a lot right so first i need to be 19:31 be an expert in the cloud provider and then i also need to look after the scripts and i’ve 19:36 written those scripts myself and sometimes they get out of hand all the time but sometimes they get out 19:42 of hand and it does happen you can end up in a bit of a messy situation and 19:47 so the thing is instead of you going through all this pain you can 19:52 lean in lean on upon appear folks you’ve got some tools that let you do this 19:59 that’s that’s that’s an excellent point that you know the the when you write a script for gcp 20:05 yeah and you write a script for aws and you write a script for azure yeah 20:10 that might very well be three separate people because they require three 20:15 separate sets of understanding now sometimes you’ll get somebody who is a true polymath and they can understand 20:22 all the different providers just fine and they’ll write scripts for it um but often it’ll be different people so you 20:28 you you have a a multiplicity of people yeah and you also have the 20:34 the other issue which i was talking about which is you know somebody to call you who are you going to call when it 20:40 breaks well you know my gcp terraform script blew up well you know what do i do now yeah um 20:47 so support’s an issue that’s cool because because another i’m guessing another thing is when you 20:54 when you write your infrastructure code it could be code right um you can create 20:59 your things for example you can create buckets uh where you store your files in and you can create the clusters which 21:05 are open to the world and you might not be aware of the security loopholes that you might have 21:11 left you know like i’ve done that i’ve left some things out there which which shouldn’t be open oh yeah somebody has 21:17 come to me and said well you we did a scan and we found out you’ve got some stuff open uh you need to sort this out 21:23 and of course there’s just so much for for us to pick right and then 21:28 that becomes also a problem and things keep changing too and you have to keep on top of what’s changed all the apis 21:34 keep deprecating you have to keep making changes everything becomes new and just about to take in right no it’s it’s Security 21:41 an awful lot taken i mean you make an excellent point which is just because you can write a script to bring up a 21:46 terraform excuse me write a terraform script to bring up a kubernetes cluster has nothing to do with whether or not 21:52 that cluster is secure and built according to best practices um as a friend of mine used to say 21:58 terraform and these other scripting technologies like cloud formation they are a thinking professional’s tool 22:05 that’s cute that’s true you really need to know what you’re doing yeah um and and you know going with the other 22:12 approach like i was mentioning which with in this case appia the the day two operational concerns are 22:19 baked in so that you don’t inadvertently uh or at least not easily inadvertently do 22:25 something dumb with your security so that you you build things that are uh insecure yeah um 22:32 or unsecure i always get those two words i think we all know what we’re talking about yeah not secure not secure 22:40 not secure thank you that’s a better way to put it um no that makes sense right so this is 22:45 what the the the devsecops what started is saying yes shift security to left so 22:51 bring it before you know like sometimes like back in the days i’m going to say back in the days security perhaps used 22:57 to be awful thought everything’s been created and you’re like ah we need to make it secure but now we’re saying with 23:03 app here how we do stuff from the get go we put in all those good practices 23:09 so of security development cycle life cycles everything is there from to start with and that’s what we do correct yeah Expertise in a Box 23:15 it’s it’s it’s really expertise in a box it’s you know we have the battle scars from doing dumb things yeah uh and at 23:22 least in my case i can’t speak everybody and then having done those things that you regret 23:27 you you and you encased or ensconced those those best practices in code 23:33 so that nobody else does it um so that that’s that’s the uh the approach that we take is that is to to 23:41 be secure now it’s one of the things that’s nice about that approach is what you do is you you would 23:46 define a user id and that user id inherits all these security hooks and as a consequence you 23:52 don’t have to worry about uh the the the individual security changes that you 23:58 would have yeah yeah no that’s that’s fair enough i think well if if people are interested you know head over to app 24:03 me at our i o you can you can find us uh on you know uh on twitter or we can give 24:08 our credentials in the end you can you can message us and you know we can we can have a chat but let me just welcome if 24:15 you’re joining in uh now we are talking about business scalability and one of the big issues that we face what that we 24:22 we work with our clients that we see is technology wise we can’t people can scale up and 24:28 what then it tends to happen is the business scalability how do we onboard people how do we make sure people get 24:34 their resources allocated to them correctly and that degree onboarding users doesn’t 24:41 become bottleneck users as in like developers and people who need to use those resources and the way not to do it 24:47 um the way not to do it is by throwing more people at it and keep doing this stuff manually and as tennis 24:53 was explaining you can automate that stuff you can automate using any scripting language or any script you can 24:58 write a bunch of scripts but sometimes you have to of course maintain them and keep on top of it but it’s probably 25:05 better to you know maybe invest in a tool that does it out of the box and we’ve got those expertise and if you’re 25:12 interested check out our website at our i o we have some some fantastic tools there we’ve got all the um all great 25:18 people in there that can we can help out we got we’ve got a few more minutes like i’m going to say about four minutes 25:24 before we jump jump into questions we’ve got a few questions tennis uh some hard ones so be prepared please 25:32 okay but before you know um i think it’s important tennis to 25:38 in my opinion like this is my opinion because i work as a developer person who creates stuff this this this feeling of empowerment 25:45 that i can join an organization tomorrow and start building things and start 25:51 deploying things to production where people can start using it i think that’s very 25:56 important to give into the hands of developers support people sre people where they might be would you would you 26:01 agree with that absolutely and that was trying to get to yeah absolutely that’s a really excellent point 26:08 we the the the intent behind the apia approach is Apia Approach 26:14 to stand the paradigm on its head instead of going and saying please sir 26:20 may i have a cluster and the guy goes just fill out a form and you send a form off into the ether 26:26 and somebody reviews it and and approves it and then a devops guy is 26:32 assigned to it and that devops person goes okay i’m going to build the cluster i’ll get back to you after lunch you 26:38 know and you know and by the end of the week it’s bill okay and then you eventually get another message out of the ether that says your cluster has 26:45 stood up then you try to log in and you can’t uh because they forgot to tell you about 26:50 some credential and so that’s the old sort of way of doing it well 26:55 the app via way is upside down for on in on purpose which is we pre-create 27:02 the permissions and we set up a user id for you based on the administrator will set up the id and 27:08 then you decide when you want the cluster and what you bring up the cluster and 27:14 immediately you’re off and running um and salmon i’m sure you’ve gone through this with with other organizations that you know i’m a brand 27:21 new developer and i’m here to develop and it takes about a week to get 27:26 an environment yeah and let’s just be conservative and say it takes it it takes like a week to get an environment 27:32 where you can actually start developing um and this this the apia approach is very different 27:38 it’s upside down which is you drive the effort to create the environment on which you want to do the 27:44 development uh and that’s the intent yeah yeah so basically we want to 27:49 accelerate that process well this is the scalability right process scalability that’s what we’re trying to get to 27:55 exactly you know they’re sitting around trying to figure out what to do and you know waiting for permissions and all this 28:01 sort of stuff and it doesn’t just have to be uh kubernetes right tennis uh 28:06 other expertise too anything to do with cloud native we’re the people to talk to 28:12 yeah kubernetes is just it is a convenient yeah it’s it’s it’s a good and and it’s 28:18 it’s a hot topic so it’s it’s well worth discussing um well so you 28:24 you do mention tennis that we shouldn’t hire people and automate right so let me 28:30 let me just summarize well you’ve scaled up your business technology scaled up and uh you need to create more resources 28:36 because you’ve got more people joining your organization and instead of manually creating them you automate them 28:42 don’t hire you know people to solve the problem you automate all these issues and you can and you can go up to an extent where you 28:48 can keep automating stuff keep writing scripts but then you have to look after them you have to maintain them you have to follow you have to make sure what you 28:54 implemented follows best practices if you go through all the documentation of everything and things change you have to update them or you can you know 29:01 speak to out here and we’ve got some tools that helps you so there’s a question that i think is a 29:07 very good question that’s come in and i’m going to ask you that is sure tennis you keep saying we should not hire more 29:12 people to solve this problem when do you need to hire more people oh good question it’s a good question yeah um Hire People 29:20 you need to hire people to do uh the automation you don’t need to hire people to do the manual stuff so if if 29:28 things are piling up and you’re getting complaints uh you’ve got two well 29:35 at least two options you you hire somebody or several somebodies 29:41 to do the automation in uh for you know your own scripting effort okay 29:46 or you you you start talking to appia and and in installing apia 29:53 but the bottom line is not to throw bodies at a near-term fix you want to throw 29:58 your effort behind something that is a permanent fix okay 30:04 and and and that’s and i guess it boils down to when you start hearing complaints when you start uh 30:10 well hopefully you can in run that a little bit by anticipating it you know if it’s if it’s clear that your your cue 30:16 last week was two requests and your folks had no trouble um dealing 30:22 with it and you realize that in six weeks it’s going to be a queue depth of 8 or 10 or 15 or 20 or maybe 50 or 60 30:30 that’s when you should start thinking about maybe i need to hire 30:36 automators instead of implementers yeah um and 30:41 and you know that’s that is not a trivial effort by itself to hire the right people but that’s that’s the the 30:48 message that you should listen to when you start seeing the q depth increase okay so when the 30:53 queued up starts to increase and the manual effort can’t keep up you should hire people yeah to automate 31:00 that stuff rather than trying to keep doing them you know assignment to manual stuff yeah you should i mean yeah 31:06 manuals the manual process should be a stop gap at at best anyway you should 31:12 you realize when you sign up for one of the one of the one of the things we haven’t really 31:17 talked about but is the the if you come from a highly 31:23 uh uh bureaucratic environment where there’s this multiple stage uh uh 31:29 vetting process and permission garnering process in order to get something done 31:35 and you and you sort of you know weld that on top of the cloud effort that doesn’t really work too well um 31:43 you’ve slowed down your cloud effort by doing that so you really need to to to 31:48 re-evaluate how you want to get things done uh and and and the fact that you’re 31:54 still doing things manually is just an indicator of that that overall uh effort that you need to think about 32:01 okay so i think that sounds uh like a like a good answer there which leads into leads very well into the next 32:08 question sure the question is it’s actually a very good question says cloud engineers know that metrics to they know what kind 32:15 of metrics they need to collect in order to scale their services right so how many people are visiting their application and they can scale up but 32:21 what metrics does the organization need to collect in order to assess if they have to scale 32:28 their business you know the process that you’re talking about well i sort of answered that in the first question 32:33 which yeah which is and and you know it’s still a good question which is it’s the it’s the number of requests that are 32:40 coming in and the and the and the crystal ball gazing you’ll have to do so yeah okay 32:45 that’s you know it’s one or two this week but it’s going to be a lot more in six weeks or eight weeks or whatever the number would be 32:51 um that’s that’s the kind of metric it’s and that’s part of the the the 32:56 devops ideal is to talk to your customers and know what their requirements are 33:02 um if if you are in a large rather siloed environment then you may 33:08 see it in the form of tickets that are coming in but if you’re in a small environment where your people talk to each other and 33:14 you get slack messages you know you you just have to understand that oh well 33:20 this is success you know when people say this this cloud thing you’re doing is really awesome you should hear alarm 33:26 bells you know when you’re sitting you can’t sit on your laurels you have to okay now 33:32 i’m going to have to start worrying about scaling this yeah so i’m guessing when you scale up 33:37 you perhaps also want to collect metrics on how long does it take for me to yes complete my tickets right how long does 33:44 it take for for a developer to get on-boarded so they can deploy some stuff that’s right 33:49 into production right so now those are good those are good measures yeah yeah yeah so some metrics are on 33:54 those so anything to do with you know if you can connect that and and as you say we can grab that from the 34:00 tickets that were raised and how long for it how long it took for us for that ticket to be open you know like we can 34:06 figure out and those metrics it doesn’t have to be a fancy grafana or prometheus right now you 34:11 can’t just grab it from from your tickets and get that make a dashboard and 34:16 you could you could do it with you know just you know scanning through your chat history on slack and oh look you know 34:23 the guy asked for it last week and it hasn’t been done yet you know that kind of thing yeah so i think that’s uh that hopefully answers that question quite 34:29 well well a couple more tennis for you sure and uh so the next one is when it 34:35 comes to tools is it worth thinking long term before you adopt them 34:40 should you think about the scalability of your tooling or wait till you actually face the need i 34:48 yeah i think the earlier the better um you know you know the the 34:55 unfortunately most of the time and you know this is human nature it’s when pain is in your life that you 35:02 take the time to deal with the pain but as as i was saying as your 35:09 organization grows hopefully because of its success then 35:14 really early on you should start to take think about how am i going to grow this and if so 35:20 what uh tack am i going to take to to address this 35:26 and so i guess the short answer to that question is the earlier the better uh you know get your head above the fray 35:32 long enough to realize that you know we’re we’re going down the path of this cloud thing everybody likes it everybody 35:39 wants it it is absolutely not going to get less in its uh 35:44 demand it’s going to get greater in its demand and we need to be uh tooling up to handle that 35:50 yeah so basically a bit of forward thinking yes into like what you need to do and i think like sometimes tennis i 35:56 see these people perhaps ponder too much on what kind of things what kind of tools they should use right yeah and 36:03 if you if you’re months down the line you still haven’t decided what tools you’re using maybe at that point you should just 36:08 start right we’ll go with this and then we can we can take stock i think it’s important to take stock perhaps every few weeks to 36:15 figure out is this tool giving us what we wanted and not be not be uh this is 36:21 i mean it works best when you’re not we when you’re scared of pivoting and saying oh this thing failed 36:27 let’s start again let’s do something something new sometimes of course you can’t do it but you know don’t be scared 36:33 of changing things halfway through oh yeah and going back to the terror going back to the terraform example 36:40 if you’re a small shop and you’ve got just a few people and you know and you’re growing okay you’ll grow to a 36:46 whopping 10 people in a couple of years then you know terraform will be fine but if you’re an enterprise 36:52 and you’re going to be drinking from a fire hose yeah uh you know you’re going to have to think about 36:59 uh how that organization is going to scale to meet that need and that’s that’s the sort of situation where where 37:06 happy is a good answer um anyway that’s nice fair enough i think i think yeah 37:11 this is this is good i’ll go let’s let’s do one last question okay tennis i know you’re quite busy you’ve got some sword 37:18 practice coming up some other problems scalability problems to solve but let me let me give you the 37:23 last last question there are there any resources things like books blogs podcasts that can discuss the 37:30 scalability issues and then perhaps can point people into the right direction any suggestions from your side i you 37:36 know to be honest with you i don’t know of anywhere 37:42 uh other than you know you and i talking now and blog posts that we did and it wasn’t just me it’s got my name on it 37:47 but it was you know it’s a it’s an effort among a lot of people to talk about this um other than that it doesn’t seem to be 37:54 as a topic that is covered a great deal i mean and and maybe i’m just talking out 37:59 of sheer ignorance if somebody wants to chime in and you know leave a comment that says otherwise but um 38:05 the the we get as and this is a human thing is it we get so 38:11 enamored with the shiny thing that is the cloud yeah that we don’t think about the 38:16 personnel implications the the process implications yeah uh as much 38:22 uh so the short answer to that question is i don’t know of any uh okay but but 38:28 i’m happy to be wrong on that point no i mean there’s some bits you can 38:33 perhaps there’s some bits not all of it and i think it’s a very valid part that you make tennis that uh 38:38 people you just talk about like technology problems and the process change i think there’s 38:44 that you can get bits of it from you know the phoenix project with the classic the classic phoenix project 38:50 great that’s it that’s a good book i think there’s some bits of it inside like in terms of processes well there’s another podcast from app yeah if you 38:57 just said rapid podcast we got a few episodes out already uh sometimes it is a topic of discussions 39:04 as to why some of the projects failed and the answer is the processes were in place so yeah uh i think if you are 39:11 people are interested you know you can follow us along on on the webinars and you know we can do it anything else you 39:16 want to add maybe i can do a quick summary and we can uh we can we can end there anything tennis you want to add 39:23 that we haven’t talked about no sir so well uh thank you all very much for joining the the webinar so the key 39:29 takeaway is that uh when you move in from 39:34 when you’re new into the cloud and everything is working great you can scale up technically correctly you 39:40 should pay attention to your process your business process how are people getting onboarded how you’re creating things at 39:46 scale and if you’re still doing things manually perhaps think about thinking doing it in 39:52 in an automated fashion used you know infrastructure as code or anything you like 39:57 but then you have to of course look after them and implement the best practices the security practices or 40:03 you can speak to us app here check this out appvia.io we’ve got lots of tools lots of people lots of discussions we’ll 40:09 be doing more of these so let us know anything you like you can find me on twitter as soul manifold because i couldn’t get 40:16 some money but it’s a song so you all mean i i tweet about like uh tech stuff 40:23 at times and i also have a youtube video which you can check out and make some uh kubernetes videos with the same name saw 40:30 my neckball so you know what to say tennis like and subscribe yeah feel free to drop a message and you 40:35 can talk to us tennis would you like to share where people can find you oh i’m just just um 40:42 tennis.smith.appbia.io oh yeah same right salman so please drop in 40:47 any questions that you have anything you want to discuss scalability technology wise and we will be very happy to answer 40:53 your questions uh we’ve got uh i’m just gonna put slide up all right so i think that’s 40:59 that’s done there so thank you all very much for joining uh and hopefully you will see on the next one 41:05 thanks a lot tennis it was a very insightful discussion thank you sir thank you all 41:10 bye-bye