Bestand wählen
Merken

How we adapted our DevOps Pipeline for Chef Delivery

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
professor coming were really excited to be here this is actually my 1st shift conference so it's more more than the for me here but I wouldn't wish it up a little bit in this session and that wouldn't do to really lighting thoughts and it's going around pipelines as well as the importance of having a shared delivery workflow especially in electors world around the support and defend the management that with about today so before where she jump
into the actual pipeline tools I want to take a step back and can revisit how that UPS has been defined over the years because that's a really important often understanding before actually looking at full that puts in the action the back in 2014 gene actually described the lot is morally philosophical movement than a set of practices I and right now certainly things have matured but in services environment we do we do consulting services rely customers we still see a lot situations where there's a lot of turning on and try and understand the what the how and the why they're not as chill as the People's picture but you can feel the tension and the excitement that there can talking around at so what what has changed recently
and a lot of provision has taken place and what we've done abuse on is basically the spilled it into some practice areas and we've been applying this over about 2 years and helping our customers organizations about DevOps and why this is actually really useful is that it gives them a real tangible and set of practice areas that they can actually but the arms around forces anymore just the debate around the delivery of the admiring things like that so having something is defined as as some like this is important to the help to refine the best practices but also help to refine the tools are available out there and most important 1 right now is run delivery pipeline so when I go to the customers it's very hard when I tried explain the but when I go and give the presentation on some practices immediately followed up with the delivery pipeline demo they start understand what that actually looks like and and that's the purpose of this is the talk is to look at the current pipelines and how it has been adapted to shift the literary this go back for
more so what what exactly is a billiard pipeline there's different definitions from how we won a baseline it is actually your dad in action so the set of practices the set of processes and tools on that you provide the workflow automation that reflects your dial-up strategy philosophy as well the practice is that you've kind of rallied around but in the session would actually do a full blown up and compare and contrast against the 4 pipelines like the city of adjectives to King city Spinnaker would highlight the key elements on from a veritable implantation which we see all time which is extension of Jenkins city and all the planning their use actually the workflow stuff and how we actually look at that and adapted to assess delivery because primarily of the Tennessee support it so hopefully this is
familiar enough elaphus here that have not quite diving to shuffle the readers some these other tools and Jenkins CI the way we extended to work it is the literary and workflow management there's a whole bunch of plug-ins in addition to a ton of grew scripts that we created taxi handle the arm movement between the 4 stages so for a specific project it's fine it works but if you can imagine if you have an environment where you moving towards might services and there's multiple the re it becomes really carried that implement the I handle you gonna handle dependencies on how that integer environment and things like that and L. Jenkins is not problem that the that its job is to execute and build analogous stuff so this is what we've got to the other is a pre pre come pipeline mom and we've seen here in terms of innovation is the bill pipeline so it does the rendering of the upstream is all that Austrian connect jobs so shows you Dupree good agree on where things are actual status and a whole bunch of other ones like extension unification and then publish your SSH which allows a build our effects that should be pushed through the SSH so fairly common
which I so what's what's the point of that so the point I'm trying across is that just because you have automation and repeatability which are really cool things does not equate to having a real good solution in terms of effectiveness and scalability and the best example and I think gosh fossil users before is your typical Rugova contraption that you this picture the incentives jobs as the get to face that that it is the automated sure I can do this all day long without having intervention is a repeatable yes I can repeat this anytime 1 but is it best for the job probably not and if you pry have to look at your pipelines the matter in terms of are you forcing tools to do something that's not for and is it worth it into the long run when you have a Marx's designed in its scale it up so that's something to think about so how do we
actually can adapt to that specific like chef delivery so this should be pretty Miller diagram of this showcases the actual delivery pipeline as defined by shaft done it is fairly opinionated and I will admit that when we initially reviewed this arm initial reaction will for customers was that was the and it was a basic because of ah entrenchment with Jenkins and lot custom work so we felt like we need to have the the flexibility that a changing around so the 2 things that really part that was done the 2 men of steps in terms of of of accepting code change as well as pushing the donor accepting changes of the actual things coming through and pushing it out on a few action was we need to modify that we want to deal with on it that Our situations and what meant that as initial reaction and the other 1 was we need to be able to mix up the different the major steps inside In theory that's not doable who just live so all we did was we took the back on and really look at the requirements and see how that action aligned and what we found was actually a pretty interesting so as we counted the back the many steps in a lot of sense and especially in the governor world which is a lot of customers in terms of their processes of having an a man a process that you say yes I reviewed this code is good an angles other 1 so now point we're able come to a consensus section really good thing in alignment the other 1 was the pipe oneself is not bad and we were able to map our different steps with the chef delivery fairly well without having major movements within the workflow so those OK and the 3rd piece which can pushed it over the edge is around the union environments and we're helping our customers going to a new multipolar restrains Mercosur's world and while the biggest challenges around how do we handle the dependency support and wise actually go through a demo that we put together to highlight how that's a really important step environment the non so that really push the edge we are able to have multiple pipelines go into the shared environment and the the control that in a really nice man so what those are doing this but when the Wang the so the very same history
somewhere that we can't really look around the steps but there is no more which is already that were in place of the customers in the built-in so we alleging that which is overlaid on and made a lot of sense as you can see you the verifiability ceptance union rehearsal we kept it but there's different done throughout but the steps inside each major phase is customizable promise the max so we had a lot of things that we had in place like you I Test modify we test mothers manual or automated and a bunch of compliance things within the environment and so that the key takeaway into the adaptation is that if you had taken the place right now by all means keep it does really good job job execution just I want the republic identify the parts the handling the the prescriptions of what is the handling the workflow move that would accept literary and you could still maintain the object is shattered allow to build and over time as you feel comfortable with that you can start using the worried actually replace on as well to have a more unified are solution so we've done this a couple times and it seems pretty pre straightforward so what are some of the key
takeaways and for the 1st part of this talk is to look at your pipeline as enterprise workflow system because if you look at it purely from a confused the re get into the mind-set of looking purely at tools looking at the actual activities and steps which makes it a lot more complicated so keep the steps simple or as and did not love this morning easy figures that public stuff but as long as easy to do is maintain at a good thing the other 1 is avoid the huge contractions and there's a lot of legacy stuff that work on that has this huge massive see ICT that it takes a lot of effort in and manpower that she maintained so start fresh try compared them i and the 3rd thing here is to use tools what the net for a lot of vendors alive tools have a plugin ecosystem audibility the customized so you use a your own discretion but don't try to bend something to well it's not really for you can pay for it down down the end of the road and the last piece is like nothing reiterate this is your pipeline as a workflow that messages looking at from the ICT so when you guys look at a for evaluating enterprise solutions solution need to apply the same evaluation criteria but how easy and how visible the having to the steps versus you know how fast some or how well it induces protected built so you have to balance that out so with that of whites jump into the the 2nd part of the talk which is going to be in it for the might of services but also are highlight of freedom of speech at the of the dependency support active looks like it will into that take it language thanks
so thanks to me so I wanna start off by in quotes you go
from 1 fell ideas both notice a if you've been reading about resources both no with his sight this great resource of course but also with the please making square is essentially that omega services are great this is kind we all want to do but they don't come without challenges in you need to plan for the challenges upfront and expect them or else you probably be kind of a world of hurt right so on you know pull this picture from wikipedia this can like fighting is evil serpents in this town like left felt like over the past year in like working on adopting marker Service Strategy for some of our clients and starting to hit some of these actual John talk about itself it is my hope of free will spread this talk in your attendance the chef is that it can go home and starts Larry Summers such as serpents and as you document URIs strategy also just like we
have a home 7 this kind of steps from deadlocked concept that what we thought about mattresses well it's by no means is like exhausted or like authoritative list and again if you folks are looking for additional resources and services and Newman has a great book I should as well but this is kind of how I like to think about it on a decomposition you saw in to like this so picture of silo on in that's kind of the 1st step that people start with when about my services how are you going to take your monoliths that you probably already have decomposed on which of those pieces are redundant which of them are you business-specific which of them are more utility based on and now once you've done that they have independently operating services this is great having a talk like something that you might use to and you integrate this by 1 method call 1 other and they actually have crossed service outside integration service discovery this kind things RTP ice right how are you going to share the description of like the exchange rate you probably want thing about your API strategy Europe and management tools are in your service discovery and all those kinds of things like your life cycle versioning others all become very important things as part of lecture architecture and in continuous actually work really well my services because as you start to you decrease the size of the application make sure that your application is not wasting resources right just putting 1 micro-services like entirely and that the inaction has like a lot of overhead so containers micro-services fit very well together because containers can run very efficiently as compared to like traditional means on secure is action really great thing for like a service perspective so if you're talking to your security engineers like you how my going to prove this by the service you probably tell them will he and actually going to allow us to focus our attention on our organizations most important assets in Lisbon most times during those things but they're not the span of time so it's not other things but not have this blanket you shared security posture against everything your model and this is in a light actually focus in on that particular service that's the most important to make sure that that's right are in HP orchestration I think this is where a lot of of your problems from crop of 2 can come talk was well but how do you deploy manage schedule in where's your environment can things where the story I like your secrets like what solutions you have for that frame so the unique I think of all those
things as you develop your lecturers architecture is something that would then spinning what's on the clients on on in of course you why do this right like if you've been developing monolithic applications for a long time like everything's great great so it's not just about I'm having like the types of scale problems that Google and Netflix and others have are that like would cause you to go to my service polymers action a lot of really good reasons to do that and you know having smaller applications are pro easier understand if you're having new developers come on board unit has been as much time and getting them up to speed on this faster to build on test about you guys but I you environments where K like could compile and like it's some coffee likely happen is and why come back like I forgot what I was doing right so hopefully this will be of benefit of lectures on if you have a if you have defects iconic problem easier to solve on you can see all the money services specific to the needs of those services right so there's a particular services getting hit hard on new skill that as a result you have greater efficiency rather than cognition pushing is big monolithic application all around and in addition when like a service fails like you guys are used to call like the job script days like graceful degradation and right like you know jobs that enabled can of each cell support among all can be applied to mark services right like if your searches down you deal because customers console like purchaser staff and that's all what matters to them like most importantly torisation flickering e-commerce isn't like that and freedom to choose acknowledges is like you know why the debated 1 like right each team can choose the right technology for the job and we can't talk about the challenge on the other side and then probably the biggest misses that we've all heard at this conference I thinking iterated everywhere is like reducing the township right like very competitive market but if you can develop something we should always through production that thing being smallish problem making it a lot easier to do that so we all can 1 be this guy like he's like relaxing sonication so light and you must be doing something right and that's because he's successfully deployed services that's my goal
also what are the challenges right like what is Martin fella talking about when is like a watch out of well the 1st thing is like handy decomposes like see you like with your monolithic application like 0 man like like Berkel splicing worse off by saying like solar days like a a array of like there's demand purchasing new mean basically talks about domain-driven design and a book that was written like for along time ago like talks about that in context so you may that it on how to seismic services was 1 3 there's well there's a term of heard kind of going around calling and service so what like what small and micro service must be in service right and they go service right maybe that's like a like point is just become like a class discipline with single class some object and that's all you got the right so I mean it's kind like Goldilocks of 3 bears were like a mild stall that the forage but there's probably something that stretch rate go figure out what that is not for your team the organization how works on how you integrate services and and as well like this to like very important aspects of this you know the problem will thing about rest but a lot of folks are also doing like just RPC greatness like Google probe this other techniques to do it interacts very hot in very performant ways between services and and but you still need to have like you're in time management strategies like internal integration versus action iteration highly interface with you I was right in mobile or other other ones and data use book your schema like what happens to your referential integrity when got like each service having its own database so that data store so you gotta think through of those problems we need like score of 4 out on data when you know for records change as compared to what you use in relation model and accordingly deployment like versioning begins with important on non-breaking changes of course of things problems as like like her a lot about using semantic versioning is wrong for aspects to like understanding and trying to coordinate across the points that's what we'll talk a little bit longer than of how we think delivery can actually help the process and how to balance the rights of technology so clients are trying to ask about like that these governance bodies and like saying well we need a standardized set of tools that are used in some ways like in time like a service right like what would you standardize that way and it's changing needs to figure out the right off the top of that same time you don't want to many tools various make it hard for developers to move from 1 project to another so also have to learn everything on so chef says
that but and you know you should use 1 pipeline provide service and I think that's actually like a really good model right you got like this for a set of stages that the action test this things independently and you bring them all together and make sure that they work in the right way so you know I think that's basically approach that resigned you know turning as well when we're looking to the voice of shelf delivery for like services and so this is just like
other really super simple on like demonstration of like a problem you can write into trying to coordinate retains are you got you know some a note renowned service b and this is like breaking change that's planned for service the in and maybe you could say that the Orion someting is for all but at some point you know version 1 is gonna be deprecated sunset in in version 2 is going to be the thing to do is to have a service a figure out like I know when they need to either change the implementation right like so maybe they're going to I still be on version 1 when version 1 instead or maybe don't push person to out when version 2 isn't ready in which environment dealing with this kind to is this all this you can try to this is more of a process approach so again management is kind of the way to do that right you got a community these years he got get token's you're watching whose using your API eyes very closely in your being a really good neighbor like when someone is consuming an old API that you're planning to turn off in reach out of these people because they have EPI teasing exactly where they are I mean it's a lot like the CPIs going off 1 is 1 warn you deprived you probably have had received those e-mails from different services that even using that's right so that 1 model but is not like a guarantee right the study we some things are out there are that like maybe some of the teachers things and like that happened so the other model will be of more of law enforcement right like actual tolling that's going again you some sense of guarantee that you I have something pushed out in and not coordinated well in I think that's were actually are shelf delivery you can make a big difference there's a capability exists all of testing management when you create a new for each project rates for for each project you can basically a dictator a specified dependency and when it gets to the Union stage it will actually run the smoke test for the other dependencies depending how you could that I can help to ensure that everything is the way that you expect it should be OK so
on tourists coming to a screenshot approach just because things were going along with slower try running this on actual results because of when using it probably say that roller stages Israel's unit always get them so what I've got arrested up is very using of
this tutorial will also hopefully you guys can come to go the Sapporo and adapted a similar way as you try to go from the application it has to move like a service for which they are so the spin of units and servers that kernel of like this but I didn't civil war you get like to build the odds of a shift server Chef Server workstation it's like like Windows box and you get like 1 acceptance no uh so I added more the standard on some works of instances and something on you know it's an all see how I cannot link this to the various stages of the pipeline so that when this when this projects are going through the different stages in be deployed into this I I created for
repositories in the customer service is like having public Austin customer service that's the 1 you find the tutorial so you can come start from there and then create additional ones this will become lectures so wedges induced comedy e-commerce product this traditionally 1 Apr now got 4 different records that will be for reference services and products customers and new orders for purchases and then selects a new wider title again the and once you and do this chap delivery in it on each of these repose universe something that looks like this which is on you know the all the all been registered in the delivery demo it
would that will also do as you begin to run them through pipelines is actually push the shift of of recipes and cookbooks to the shelf server itself but what I did on what I got to that stage was actually looked at so actually so the cookbooks would be pushed as well as of course always needs that use all right so rich through literature and is basically configured for the particular environment so we run through the of particular phase shift delivery it sends the environments to ships server and says Ron any cookbooks for this environment for this project of so consular works but I just basically when in breach of acceptance ones but should deliver automatic rich environment for acceptance it creates 1 as you can see here
a create creates 1 union so
all of the union servers or docker containers or whatever using for our services I just like to that it I modified occurred and so
this is like the candy del box when you guys use of the tutorial of course can change it right so what I forked It's a products and it like ones I'm interested in like X boxes and of and then like a purchase would just be looks enjoying it needs to sing like relational world right like some foreign keys so like that so obviously this can be dependency between the purchasing and other services so if you look at
of the tutorial they've like example was smoke test in all this basically does is it iterates through the nudes that it finds himself server for the given environment and it calls the EPI may choose to enter a k so like what needs to be done here is actually do API test right you need to test to make sure that everything is fine for the EPI you're expecting to exist for yourself also for others right and what will happen is when we get to the dependency part they will run the smoke test for the other servers to so that you this is basically where you can guarantee that you have the guy in the environment you expect otherwise failure smoke test to make sense so dependencies of other services that work is also dependencies that you have and others so is like 2 different ways you can do this and this is how you basically on we describe a dependencies use the actual name in chapter delivery and inside a project so remember for project is going modify that but then you
run it through like the you'd Overview commands gives the pipelines
so you can see the smoke test
in the cool thing is that like just call that could serve in the army but 1 is the
acceptance phase just 91 right because there's 1 environment for this form are mistreated acceptances rates so there's only 1 thing that's going to return people go to the union it's actually going to do
for because it found 4
boxes in the union environment that's just the way the smoke test is written you can change the the way this practices developed 2 different things
induced what happened when I ran the customer service so not only did you run the smoke test for the order service like efforts in service but it also renders smoke that's for the water service also you you basically have to your direct dependency directions correct so you in nature that of the you water depends on customer then or customer to order then customers or service vice-versa but you know all you have to do is add an entry in it will actually start to do this I want to get the C union stage also that's basically it so
hopefully like when you guys get through that inter and intra think about how to apply that adult that 2 lectures strategy and this will be helpful resource for you to both validate held that is working really expecting to and also the think about how you want transition that today a service model so what are the takeaways on where services are also do it but just like plan ahead right like you like right so that things I should debris is definitely helpful for and managing that versus trying to coordinate and more process origin way in they should merge services should probably not be and after 5 so there's a lot of debate about like to start a monolith thing there lecturers but at least for each marker service you probably need to think about I wanted that is like at least for our clients we've had things like like services for 1 big relational database were like now trying to basically split that relational database as well so we I went and thinking let's put up the database and services at the same time many depend on the off at the end of the day so that said i thanks folks so thanks
Bit
Wechselsprung
Verschiebungsoperator
Demo <Programm>
Dienst <Informatik>
Forcing
Flächeninhalt
Menge
Selbst organisierendes System
Gruppenoperation
Kombinatorische Gruppentheorie
Programmierumgebung
Hilfesystem
Prozess <Physik>
Gruppenoperation
Gebäude <Mathematik>
Term
Jenkins CI
Eins
Homepage
Datenmanagement
Rendering
Prozess <Informatik>
Code
Skript <Programm>
Maßerweiterung
Soundverarbeitung
Umwandlungsenthalpie
Addition
Prozess <Informatik>
Stichprobe
Elektronischer Datenaustausch
Plug in
Gruppenoperation
Dienst <Informatik>
Menge
Ganze Zahl
Strategisches Spiel
Projektive Ebene
Programmierumgebung
Umwandlungsenthalpie
Soundverarbeitung
Zentrische Streckung
Demo <Programm>
Punkt
Prozess <Physik>
Winkel
Skalierbarkeit
Gruppenoperation
Mathematisierung
Güte der Anpassung
Term
Ranking
Physikalische Theorie
Code
Spezialrechner
Diagramm
Skalierbarkeit
Prozess <Informatik>
Reelle Zahl
Mereologie
Garbentheorie
Programmierumgebung
Metropolitan area network
Softwaretest
Perspektive
Gewicht <Mathematik>
Schlüsselverwaltung
Formale Sprache
Güte der Anpassung
Programmverifikation
Datenmanagement
Sprachsynthese
Plug in
Diskrete Gruppe
Physikalisches System
Design by Contract
Unternehmensarchitektur
Objekt <Kategorie>
Physikalisches System
Dienst <Informatik>
Prozess <Informatik>
Anpassung <Mathematik>
Mereologie
Figurierte Zahl
Schlüsselverwaltung
Programmierumgebung
Unternehmensarchitektur
Message-Passing
Phasenumwandlung
Leistungsbewertung
Auswahlaxiom
Dienst <Informatik>
Client
Umwandlungsenthalpie
Architektur <Informatik>
Quadratzahl
Rechter Winkel
Strategisches Spiel
Kontextbezogenes System
Resultante
Überschall
Rahmenproblem
Selbst organisierendes System
Stab
Gruppenoperation
Zellularer Automat
Kartesische Koordinaten
Whiteboard
Fehlertoleranz
IntServ
Client
Informationsmodellierung
Einheit <Mathematik>
Datenmanagement
Maßstab
Prozess <Informatik>
Perspektive
Code
Datentyp
Skript <Programm>
Softwareentwickler
Softwaretest
Zentrische Streckung
Videospiel
Addition
Computersicherheit
Softwarewerkzeug
Systemaufruf
Mailing-Liste
Biprodukt
Bildschirmsymbol
Bitrate
EINKAUF <Programm>
Arithmetisches Mittel
Helmholtz-Zerlegung
Dienst <Informatik>
Informationsverarbeitung
Rechter Winkel
Mereologie
Dreiecksfreier Graph
Strategisches Spiel
Computerarchitektur
Overhead <Kommunikationstechnik>
Programmierumgebung
Bit
Prozess <Physik>
Punkt
Selbst organisierendes System
Desintegration <Mathematik>
Klasse <Mathematik>
Gruppenoperation
Mathematisierung
Versionsverwaltung
Iteration
Kartesische Koordinaten
Dienst <Informatik>
Term
Eins
Client
Informationsmodellierung
Datensatz
Datenmanagement
Maßstab
Code
Ortszeit
Softwareentwickler
Speicher <Informatik>
Figurierte Zahl
Metropolitan area network
Datenhaltung
Relativitätstheorie
Einfache Genauigkeit
Kontextbezogenes System
Bitrate
Integral
Objekt <Kategorie>
Helmholtz-Zerlegung
Dienst <Informatik>
Menge
Rechter Winkel
Strategisches Spiel
Projektive Ebene
Standardabweichung
Subtraktion
Prozess <Physik>
Punkt
Desintegration <Mathematik>
Gruppenoperation
Mathematisierung
Versionsverwaltung
Automatische Handlungsplanung
Datenmanagement
Implementierung
Dienst <Informatik>
Informationsmodellierung
Datenmanagement
Vererbungshierarchie
Figurierte Zahl
E-Mail
Softwaretest
Beobachtungsstudie
Objektverfolgung
Güte der Anpassung
Token-Ring
Programmierumgebung
Bitrate
Dienst <Informatik>
Menge
Rechter Winkel
Projektive Ebene
Versionsverwaltung
Programmierumgebung
Streuungsdiagramm
Dualitätstheorie
Quader
IRIS-T
Kartesische Koordinaten
Binder <Informatik>
Teilmenge
Dienst <Informatik>
Einheit <Mathematik>
Bildschirmfenster
Arbeitsplatzcomputer
Server
Skript <Programm>
Projektive Ebene
Brennen <Datenverarbeitung>
Verschiebungsoperator
Instantiierung
Standardabweichung
Demo <Programm>
Dokumentenserver
Biprodukt
EINKAUF <Programm>
Eins
Teilmenge
Dienst <Informatik>
Datensatz
Keilförmige Anordnung
Server
Projektive Ebene
Ordnung <Mathematik>
Grundraum
Programmierumgebung
Demo <Programm>
Verschiebungsoperator
Dienst <Informatik>
Quader
Rechter Winkel
Relativitätstheorie
Server
Biprodukt
Schlüsselverwaltung
EINKAUF <Programm>
Eins
Inklusion <Mathematik>
Softwaretest
Dienst <Informatik>
Verzeichnisdienst
Systemverwaltung
Rechter Winkel
Mereologie
Server
Projektive Ebene
Versionsverwaltung
Programmierumgebung
Einfügungsdämpfung
Teilmenge
Softwaretest
Bildschirmmaske
Rechter Winkel
Client
Bitrate
Programmierumgebung
Phasenumwandlung
Softwaretest
Dienst <Informatik>
Quader
Konvexe Hülle
Wasserdampftafel
Client
Ordnung <Mathematik>
Intelligentes Netz
Programmierumgebung
Richtung
Relationale Datenbank
Client
Informationsmodellierung
Dienst <Informatik>
Prozess <Physik>
Adressraum
Datenhaltung
Gruppenoperation
Strategisches Spiel
Validität
Automatische Handlungsplanung
Versionsverwaltung

Metadaten

Formale Metadaten

Titel How we adapted our DevOps Pipeline for Chef Delivery
Serientitel ChefConf 2016
Autor Chaffee, Wyatt
Pham, Jimmy
Lizenz CC-Namensnennung - Weitergabe unter gleichen Bedingungen 3.0 Unported:
Sie dürfen das Werk bzw. den Inhalt zu jedem legalen und nicht-kommerziellen Zweck nutzen, verändern und in unveränderter oder veränderter Form vervielfältigen, verbreiten und öffentlich zugänglich machen, sofern Sie den Namen des Autors/Rechteinhabers in der von ihm festgelegten Weise nennen und das Werk bzw. diesen Inhalt auch in veränderter Form nur unter den Bedingungen dieser Lizenz weitergeben.
DOI 10.5446/34635
Herausgeber Confreaks, LLC
Erscheinungsjahr 2016
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract This session provides a look into Chef Delivery and how the pipeline is used and adapted across development teams. The shape of your delivery pipeline is critical as it controls your DevOps workflow. In this session we will demonstrate how Booz Allen leverages Chef Delivery and our lessons learned in capturing, developing, and deploying software for a federal government agency that drove us to adopt Chef Delivery. We will also discuss various Microservices design and deployment strategies that can be applied using Chef Delivery which will provide the pros and cons serving as a reference point for your current and future efforts. This enables organizations to have multiple feature teams working on different aspects of applications while ensuring confidence that changes deployed are controlled, tested, compliant, and expected.

Ähnliche Filme

Loading...
Feedback