Merken

Protect your users with Circuit Breakers

Zitierlink des Filmsegments
Embed Code

Automatisierte Medienanalyse

Beta
Erkannte Entitäten
Sprachtranskript
um right now we're going to listen to circuit breakers by Scott alright who thinks everybody Veronica now so let's talk about circuit breakers today before I get
started on a given a quick shadow to yell at Lincoln was kind enough to send me here if you don't know and you know it is a website where you can discover a small local businesses where we happen to be we've got a both outside and we happy talk to you more about what it
is that we do were pretty big website as well so in particular regarding 19 million monthly users over 100 million reviews which you know something that a lot of our code deals with and launched in 32 countries including a lot here as well as for myself I spend my time working with revenue infrastructure so that kind of code that helps you purchase products on Yelp if you're a consumer like you me or business owners well and if you are inclined toward you can find me at actually alright so if all
that away with go and talk about some circuit breakers In the old
days we used to have this very nice model of the world right pretty much any code that we wanted to call whether it was something simple something complex was executing in the same type of process that we are calling from so if I wanted something nice like display of businesses name in a nicely formatted way I call a function to do it and I can be pretty sure that functions can execute conducive
cleaning unfortunately more and more these days that assumption that code is running locally code is executing and processes not becoming so a lot of different trends are causing is things like Dr. we may have heard of micro-services or service-oriented architecture movements in general are introducing more more network boundaries and work so that function that used to be executing locally in in our same interpreter is now maybe on some remote server running elsewhere we don't even know where it is and the unfortunate
reality this is that can introduce a lot of problems right probably being used to exist so things that we assume would always work all of a sudden a strange failure modes maybe I didn't a 500 from some sort of HTTP API maybe I just see a lot of slowness and I have no explanation of what's going on there a lot of ways that this can go wrong
and that is essentially were circuit breakers fit so circuit breakers of these components that we build in between ourselves and our remote dependencies and essentially there to do 2 things they need to detect when the system is unhealthy when they can communicate to that external API whatever it is and they need to do something about that so often this is something like blocking requests ordering of features or so where are we here
circuit-breakers mentioned you like likely again see references to this book by Michael my god I release it has a lot of ideas in in a lot of ways to build the software and 1 of them in fact 2 pages of an otherwise fairly large book is dedicated to this idea of sort of so we have goal of the 1st off and introduce my god circuit breaker gives an idea of what it does but the basic idea is and why we should care at all what problem solved for us and then we're going try and walk through a variety of ways that we can take that basic circuit breaker and extended to solve different problems I will sort of motivate this by talking to a few scenarios getting an idea of ways that that initial circuit breakers just insufficient and what we can do about and do all this
need 1st introduce you to my own favorite restaurant back home in San Francisco which is just so
this is a very old school of American Diner experience so this means a vital boost sit down tables in you wages that come and take your order formula and take it to a back
and it's that kind of process of ordering food that we're gonna use as a stand-in for a lot of different things that you might want to do with remote but so is briefly what during experience distinct if it doesn't translate countries very well us to start out with many essentially you'll sit down at a table in the given a list of items you might order step 2 you're going to go ahead and decide what it is that you want the communicate that was the waiter and they're gonna write it down and it take will take the stick it and give it to the kitchen and say please make the source step 3 is can be hopefully the kitchen processing Mason efficiently and step 4 is of course whether the customer is gets their food back and then they get to enjoy so the
fundamental rule of any of these kind of modern network-oriented architectures is you have to accept that your system is going to fail and that may not be your code in particular but it turns out that because you depend on so many other things many of them may be significantly on the control of something is going and it's not a question of whether you can avoid the failure entirely it's a question of when it does occur where you can do about so let's look back at that example eating so regions and see how well failure might factors so that have order food we had a way to take down a ticket and placed in front of the kitchen cooks something unfortunate
might right so that we don't really understand maybe we aren't even aware of it but that water that was supposed to get processed and timely manner by the kitchen and is instead entirely forgotten now in a real restaurant you might hope that reaction to this would be something reasonable something like a waiter coming in telling you the kitchen is far behind that they need to take more time for your order but in the code that we actually right in practice but we have very little thought this to you right we
might not think that at all so it turns out that the real result of doing this can be really negative in fact it can be completely preposterous and if we ever send it out loud and say well that's really that's so essentially are goal today is to make sure that aware of this problem and make sure that we built in systems that can handle it on and the 1st we
didn't talk about is migrants basic search so
here's a slightly idealized schematic of gyms restaurant this is approximately true all the wife designed this way in a moment on the left there we have a number of sections each section contained in several diners in it but most importantly a section is served by a single weight so they're going to come to the individual diners and collect those orders we talked about earlier they taken across to the right side of the diagram and drop them off to the kitchen and they're a bunch of cooks in the kitchen and as they become free they're gonna pick up the next order cook preparing and handed back to the leaders to bring back customers and I said earlier that this is sort of a general standard for a number of things and so I want to briefly convince you that this is a generic kind of interaction model here's what it might look like if instead of you know the customers in a restaurant we're talking about services back and services and some sort of HTTP based API and likewise if you're dealing with task use or any sort of Slovak and process with workers you can fit that in this model as well so all of that so we're going to focus on the dining each case and what were essentially do is build and circuit breakers to this model so the ones that will start out with the going beyond each week individual it's everywhere will have their own circuit breaker paying attention to the
success or failure of the workers and in migrants basic model there are 3 essential states now on the right side of this these are the traditional names that gives to the states but I don't know about you all that I find them horrifically confusing and that I'm not used so if you're curious to you can look to the right side of this diagram but if you're a mere mortal like me I I prefer the ones on the left and on to so healthy is good state as as you might imagine essentially all the requester flowing successfully the circuit breakers can recognize that things are going well and declare itself unhealthy is the exact opposite of course if no requester succeeding whatsoever we're going to declare the system unhealthy and hopefully take some corrective action and United circuit breaker has an intermediate state where we maybe quite sure whether we're good or bad and want to decide between
2 so a very basic circuit breaker in Python could look something like that so each waiter's gonna start out by asking themselves do I believe the systems healthy right In good case where this whole thing they're going go ahead and just send the request on the kitchen people take your order and I'll make they pay attention whether it succeeds In that case the bottom we're going to go ahead and just block the request of so go and read my orders a customer I said I would really like to have uh maybe bacon a and the waiter's gonna say I'm sorry kitchen is unhealthy can't accept and we might be in the middle case that we talked about of recovering trying to decide when the
system if we are in that middle case so again the basic magnets can break says that we should be taking this kind of approach to start by reading a certain amount of seconds maybe 1 2nd maybe 5 seconds this is the given back system hopefully some sort of time to recover and then the very next request that comes through we're going to pass it to the back and we're going to see if it's successful and if so declare ourselves healthy and see if it's unsuccessful and it's ourselves unhealthy and kind of repeat this cycle until we have become healthier so that's the basic circuit breakers not terribly complicated political and explain why it's actually doing us any good for so before we answer could break directives 1 effective kitchen slowing down was the dimensions start with a very long time to get their food so we use our skeleton from earlier in extraordinary long time in the way that most of us often right this code in practice there may not be a kind of any sort of clients they just wait forever on the other hand the kitchen the back and is also getting a growing backlog right maybe they normally have 5 orders and process and as they get slower and slower than 10 they have 30 100 and the hopelessly and every dynamics entering a restaurant is actively making the situation worse so maybe in a normal operation we expect roughly 10 people died in the same time but everything starts slowing down this way we can get 20 in the restaurant what's a hundred people in the restaurant once kind of compounding
obviously problems in a circle but does make this that so for starters we actually you're frustrated users what might be a little counterintuitive given that were blocking the request immediately but it turns out that in most use cases that actually preferable the weighting knows how long and then being told information we're introduce load on the back end so the kitchen rather than having the pilot can get a little bit of a breather and recover whatever the problem was and maybe most importantly we have a place in code or we can define a failure mode 1st so you can say if the this is unhealthy in some way I know exactly what's going to happen and it's in this part of it turns out that that's actually really valuable as we start to expand and make this more
complicated so that is the basic circuit breaker I can 1 point of view assumptions were making and will use that to launch into more detailed discussion discussions so the 1st thing we're assuming is that all those readers and independent circuit breakers right and that introduces the interesting problem that 1 liter can believe the system unhealthy all the other way there is completely clueless and happily sending worsted kitchen we might think to ourselves that a little bit strange and we'll talk about what we did In addition to the basic circuit breaker we have exactly 1 thing we do in the search when the system is unhealthy and that is we stop feature requests right no other reactions I will talk a bit about what else we can do we also made the decision that the circuit breaker component itself the wages in our case was the only thing that could decide if something was healthier at that a little bit limiting you might imagine that say cook in the kitchen has a pretty good perspective on sole possible that about pulling that information and last but not least of all recovery system in the basic circuit breakers very defined and it's very focused on the success or failure of 1 particular sold out by by expanding so let's start off with the
1st question which is optional these wages have independent circuit breakers maybe should there be a little more communication between and
said originally that 1 of the weird side effects of that is they can disagree right if the weight of the section 1 is well aware of system unhealthy they won't be sending anymore request to the kitchen and that's as good a good choice for offices from reasons that we talked about previously this course the weight for section 2 men I have had any were recently and the next 5 people that come through section you're going to get very slow very unsuccessful so the
obvious solution in this case is we need to have them communicate some right and 1 easy way to do that is to just take whatever state it was that they were maintaining be the number of successful orders in the number of failed orders and put it into a shared this data store can be any number of things that my sequel something more he value-oriented Cassandra biology b but the goal is that they're communicate together come to some roughly shared consensus on the state of the system and that will be what they all of pulled from rather than their own opinion so
the new behavior we introduced obviously all of these clients these wages and having the communicating were before they were completely independent of each other and that has a lot of different sort of side effects that you may not be intention and obviously the thing that we get out of this is we're going propagate any failures in the system a lot faster so rather than every section independently having to rediscover the fact that the kitchen is really unhealthy built teach each other about it this is a on the downside this pulls in a lot of complicated questions around building a distributed data store in the 1st place and this is a short talk and I have nowhere near enough time to dig and all the complexities here but a lot of the issues around the CAP theorem in deciding whether you want to be highly consistent not all get pulled initiated shares fully so let's talk a little bit about sort of the 2nd question we had proposed which is being given to the system becomes unhealthy what we do in recall
that in the basic circuit breaker the all the action taken is we're going to block future requests now you might imagine that this feels a little bit of substandard because if I a customer I come I sit down and look at a menu I take my time I decide what I want and then the very last minute I get told were totally unavailable it's it's functional but it's pretty frustrating for all your
users so we're going to take that method that we discussed sort of deciding if the system is healthy or not and maybe that's coming from 1 later on it's coming from a central database of some sort and we're making public making API whatever the current wants to consult
and so instead of only the action we're taking being to block orders right when they're maybe we can maybe make some slight improvements like we can say don't let anybody else sit down at a table once the system is we can also imagine building a lot of other features along the same thing so the new behavior we can introduce is we've essentially made the information about helping this public and we can build any number of features on top of that maybe that looks like shutting off access to whatever this feature is when the system becomes unhealthy that made even looks like a really automatic monitoring so we can go ahead and have some sort of you know manager be notified immediately when the system becomes unhealthy and the
dangers we got here is we obviously have a lot of flexibility we can build out in the future we would like on top of this the only downside is we have to ask ourselves some her questions about consistency that information and that sort gets back to the questions we asked in previous section not so
1 other question we might say is is we said before the leaders are in a slightly awkward position to decide whether everything itself right all they can see is that the request that they make and whether or not they come and we might imagine that stands in our scenario something on the back and the cooks might have a
really good perspective here so we can have them implement some relatively simple function and they can say you know what were in the best position to tell the system is unhealthy if we have way too much work to do and we know how closely behind so we know and have them determine this and send a signal to the circuit breaker even though they don't know that says look the system is
very so pages we've introduced if we allow this we can basically say anything in our broader system is allowed to communicate information itself into the circuit and we have to come up with some way to combine if we're having wages disagree with the kitchen about whether the system is healthy we have to make some decisions and we need to decide who gets to win in that discipline so an upside we've introduced a whole world of new signals This is a very wide-ranging and powerful tool to have and on the downside we've created a tool that really powerful complicated so I've personally written some circuit breaker logic that is really really hard to untangle after the fact and I've seen something production in situations where something is on fire and I have no idea why can't we recommend that approach it's a keeping this simple enough you can understand that wall complex enough that it actually does the job is to and for the 1st question we ask ourselves is whether some alternatives to recover so in the
original formulation we had said that we're going we a specific time maybe 5 seconds and that would initially single request and if it works great were healthy and if it fails great 1 of the this research trusting that single-trial requests tells the right information 1 alternative we could take on is to try something called dark . requests and essentially what these are the bird and take my visa requests going block them just like we would have the normal circuit breaker but instead of blocking them and forgetting about the request entirely we're actually going to pass through the back with a run it through just like a normal request would be and we're pay attention whether it succeeds or fails just like a normal request now you can imagine that this is really really nice when it works we have like user traffic with all the peculiarities and features that that might have and it's gonna tell us exactly when our system is alive or were in trouble but on the downside it's thoroughly incompatible with certain systems so if I'm running a search engine I don't mind doing this I'm happy to tell the user the search isn't getting process but they actually do it but on the flipside I work with credit card processing and I think you would hope that if I told you your order had been completed I would not in the background the silently charging credit card and so they're playing cases really kind of side effects you really don't want to be telling users that things are a certain way when 1 alternative for those kinds of situations is what's called synthetic graphs so as before we start launching the blocked real user requests and then we secretly did them in the background what we're going to to do here is still what the real user request and then instead of processing those working process something I think that we've made up ourselves In the case of the diner you might imagine that every what's say 5 minutes you ask the kitchen to make a solid this is a fake order nodes can eat the salad might be more wasteful but you can go and figure out if the cell into getting made successfully than the kitchen is probably doing the obvious downside here is there's no guarantee that those that request represent real user traffic or patterns or load or demands on your system in the case that were building they're asking kitchen make salads every 5 minutes if the grill is completely destroyed and they can make a single hamburger obviously the salad requested in work just fine everything's going to report the kitchen is doing very well and the moment that you decide to send a real user request and try make a hamburger it'll break so thin that it is easy to do but maybe not
and these new behaviors and gives a lot of opportunities for traffic to determine how they let us
really move those timeouts and not worry about turning them a creating them but on the downside we've talked a couple reasons why maybe they're not appropriate for every use cases we need to think a little bit before applying art so let's wrap up to the but if you only take 2 lessons away from this talk but hopefully the 1st point is that if you have networks in your system if you're doing services a doctor any number of other technologies you need to plan for failure and not only do you need to plan you should not be afraid of taking the basic solution the basic circuit breaker if you don't have time to build anything else it is a limited approach but it does work and for the reasons we talked about in the beginning it's a lot better than doing nothing at all that said if you do have some extra time and if you can devote is there are a lot of ways that you can customize it to be even better for your personal situation we talked about maybe how many circuit breakers you should have to be 1 1 per cent of processes do you want the communication they be completely independent of each other we asked me maybe what should we do in response to this is becoming and how do we wanna go ahead and just on the west do you want a to trigger some of the features that we want to prevent users from every even seen this feature we wondered out loud maybe if the circuit breakers themselves aren't always in the best position to decide if the system itself here and then we might need to integrate multiple sources of information and make some final J and then finally we talked a little bit about alternative ways to do recovery of all the traffic based and some of the more accurate than others and I should say uh there's a lot more here obviously real-time constrained but there are many many ways you can customize circuit breakers more a pack of thinking the manner specified implementation and everything comes down use case so depending on what you're looking for a lot of different forms of this kind of approach might be the right answer for you hopefully this giving you some tools and ways to think about it that you'll find useful going
forward and with that I think we have a little bit of time for questions if I don't get any time now to talk to you happy to chat after online world the of the young thanks and will talk things and I have a question uh I've seen a lot of votes in this it is written in this so that the most of the implementation so in in the it Italy the close we have some Mason Fellow the of getting requests and just wanted to know if you do that you know you that said C so that most of the circuit breakers use our search in they make it with ah yes see certainly can build them into any part of the system I think it's very very common for them to live either in the clients themselves or maybe somewhere between clients and there's a gas but I would say all of use cases I have personally had they've lived pretty close to clients but I think there's nothing fundamental that requires them to be there that's certainly 1 of those questions that's kind of India much more traction right and interesting thing to think about what you get by putting it in clients or on the server side or somewhere in the mean in a smart proxy sounds very interesting behavior talk at the after about and I think in in a production a system where you will have probably 10 or 20 kittens so that if you will become even more and more complicated the you compare using this kind of uh it's the the the request for what the communication layer you have we simply break could you could you compare them with having now something like driving and you also queuing system should be coupled 1 thing what you write your experience to remove a great part of this complexity but maybe another 1 of their complexity arises sure shares so sort of if we have a system that's not may be strictly HTTP based rabbit and he had a number of task used to be is there any point still doing this but I can say that from my own work a lot of what we do is not HTTP based because of the nature of dealing with money is that you would rather than a network let's not forget about the order of the year competition were mostly cue-based so we do a lot of work with Amazon's SQS other technologies in that vein and even in those cases you still want to stay here right that that solves some sources of unreliability but if I write bad code in the back and that silently drops every request we still want identify that and take corrective action even if the only thing we do is on future and so better transporters are better they solve some problems but still relevant it's a great talk say that I need to my wages to synchronize over and storage of but what I do is I introduce another moving part right because this is the story store or it is another thing that could fail possible how do we you think is that what what's the best way to deal with that because if that fails I cannot even determine whether anything healthy the yeah absolutely and and that's a great observation when I had originally I talked about this with some colleagues I had taken a kind of as as an assumption that you have some sample data store because why would you want this to be independent and a colleague of mine on sort of the operational side of things basically said exactly we're describing if your goal is for this to be a bulletproof failure mode so this is maybe the traffic to search the entire web site any amount of coupling to another failure point is a bad thing right and the amount of coupling across processes might even be a bad thing and so they can be as where you'd actually accept slower response in exchange for a simplest yeah and that's that's a very good way of saying it it I see 1 more question thank thank straight up and question how would you would you reconcile this with load balancing like a government tuple kitchens and you wanna maybe you 1 1 kitchen goes long load-balanced other ones and how the make the circuit breaker track about yeah for sure but I think in some sense they can be decoupled if your circuit breakers oriented on the client side of things you may be are completely unaware of of balancing that's something that's secretly implemented on the back and died in cases where load balancer you're doing their job detecting and the instances and moving them to claim that looks just like everything is going well so in those cases the circuit breaker would not notice a problem because there isn't a problem but it's a good sort of backup mode again there are many plenty of other cases were maybe you love balances overeager they remove everything from the back and and now you're going be really excited about having a circuit breaker and placed OK I think there was a thank you very much
Web Site
Digitaltechnik
Abschattung
Speicherbereichsnetzwerk
Baum <Mathematik>
Computeranimation
Metropolitan area network
Web Site
Digitaltechnik
Biprodukt
Baum <Mathematik>
Code
EINKAUF <Programm>
Computeranimation
Interpretierer
Lineares Funktional
Subtraktion
Prozess <Physik>
Datennetz
Datensichtgerät
Komplex <Algebra>
Code
Computeranimation
Randwert
Informationsmodellierung
Twitter <Softwareplattform>
Rechter Winkel
Datentyp
Server
Baum <Mathematik>
Serviceorientierte Architektur
ATM
Digitaltechnik
Zusammenhängender Graph
Physikalisches System
Chatbot
Quick-Sort
Computeranimation
Softwaretest
Modul <Datentyp>
Software
Digitaltechnik
Grundsätze ordnungsmäßiger Datenverarbeitung
Quick-Sort
Computeranimation
Homepage
Varietät <Mathematik>
Subtraktion
Prozess <Physik>
Mailing-Liste
Quellcode
Ordnung <Mathematik>
Computeranimation
Ausdruck <Logik>
Tabelle <Informatik>
Systemprogrammierung
Fundamentalsatz der Algebra
Wasserdampftafel
Gamecontroller
Schlussregel
Computerarchitektur
Physikalisches System
Ordnung <Mathematik>
Code
Teilbarkeit
Computeranimation
Endogene Variable
Resultante
Physikalisches System
Computeranimation
Offene Menge
Prozess <Physik>
Momentenproblem
Freeware
Gruppenoperation
Zahlenbereich
Interaktives Fernsehen
Ideal <Mathematik>
Aggregatzustand
Quick-Sort
Computeranimation
Wiederherstellung <Informatik>
Eins
Task
Generizität
Informationsmodellierung
Dienst <Informatik>
Digitaltechnik
Garbentheorie
Ordnung <Mathematik>
Aggregatzustand
Standardabweichung
Prozess <Physik>
Diskretes System
Hausdorff-Dimension
Zwei
Güte der Anpassung
Subnormaler Operator
Physikalisches System
Code
Quick-Sort
Computeranimation
RFID
Wiederherstellung <Informatik>
Richtung
Teilmenge
Metropolitan area network
Client
Skeleton <Programmierung>
Garbentheorie
Rechter Winkel
Dreiecksfreier Graph
Digitaltechnik
Ordnung <Mathematik>
ATM
Addition
Bit
Gewicht <Mathematik>
Kreisfläche
Sichtenkonzept
Punkt
Stochastische Abhängigkeit
Physikalisches System
Code
Computeranimation
Entscheidungstheorie
Last
Last
Front-End <Software>
Front-End <Software>
ATM
Digitaltechnik
Mereologie
Wiederherstellung <Informatik>
Zusammenhängender Graph
Information
Soundverarbeitung
Gewicht <Mathematik>
Stochastische Abhängigkeit
Güte der Anpassung
Digitaltechnik
Garbentheorie
Physikalisches System
Auswahlaxiom
Computeranimation
Modul
Office-Paket
Soundverarbeitung
Subtraktion
Bit
Gemeinsamer Speicher
Zahlenbereich
Kraft
Fortsetzung <Mathematik>
Physikalisches System
Komplex <Algebra>
Widerspruchsfreiheit
Quick-Sort
Computeranimation
Endogene Variable
Kugelkappe
Client
Rechter Winkel
Theorem
Client
Garbentheorie
Ordnung <Mathematik>
Speicher <Informatik>
Stochastische Abhängigkeit
Modul
Aggregatzustand
Metropolitan area network
Lineares Funktional
Bit
Datenhaltung
Gruppenoperation
Digitaltechnik
Physikalisches System
Quick-Sort
Computeranimation
Software Engineering
Gruppenoperation
Zahlenbereich
Physikalisches System
Quick-Sort
Computeranimation
Einfache Genauigkeit
Portscanner
Datenmanagement
Code
Garbentheorie
Information
Ordnung <Mathematik>
Widerspruchsfreiheit
Tabelle <Informatik>
Metropolitan area network
Lineares Funktional
Ortsoperator
Perspektive
Digitaltechnik
Physikalisches System
Computeranimation
Modul
Prozess <Physik>
Momentenproblem
Zellularer Automat
Ungerichteter Graph
Komplex <Algebra>
Mathematische Logik
Computeranimation
Homepage
Physikalisches System
Knotenmenge
Prozess <Informatik>
Reelle Zahl
Suchmaschine
Mustersprache
Äußere Algebra eines Moduls
Kontrollstruktur
Normalvektor
Soundverarbeitung
Zwei
Physikalisches System
p-Block
Biprodukt
Hoax
Chipkarte
Entscheidungstheorie
Last
Digitaltechnik
Information
Ordnung <Mathematik>
Normalvektor
Telekommunikation
Bit
Subtraktion
Prozess <Physik>
Punkt
Ortsoperator
Implementierung
Zahlenbereich
Oval
Computeranimation
Wiederherstellung <Informatik>
Physikalisches System
Multiplikation
Bildschirmmaske
Endogene Variable
Äußere Algebra eines Moduls
Ideal <Mathematik>
Datennetz
Physikalisches System
Quellcode
Integral
Dienst <Informatik>
Digitaltechnik
Wiederherstellung <Informatik>
Information
Abstimmung <Frequenz>
Bit
Punkt
Prozess <Physik>
Gemeinsamer Speicher
Natürliche Zahl
Komplex <Algebra>
Datensicherung
Computeranimation
Eins
Client
Prozess <Informatik>
Kontrollstruktur
ATM
Nichtlinearer Operator
Datennetz
Güte der Anpassung
Quellcode
Biprodukt
Arithmetisches Mittel
Rechter Winkel
Server
Ordnung <Mathematik>
Normalspannung
Instantiierung
Proxy Server
Telekommunikation
Web Site
Gruppenoperation
n-Tupel
Abgeschlossene Menge
Zahlenbereich
Implementierung
Transportproblem
Code
Lastteilung
Data Mining
Task
Weg <Topologie>
Endogene Variable
Stichprobenumfang
Luenberger-Beobachter
Speicher <Informatik>
Ganze Funktion
Fundamentalsatz der Algebra
Physikalisches System
Quick-Sort
Summengleichung
Last
Digitaltechnik
Mereologie
Baum <Mathematik>

Metadaten

Formale Metadaten

Titel Protect your users with Circuit Breakers
Serientitel EuroPython 2016
Teil 37
Anzahl der Teile 169
Autor Triglia, Scott
Lizenz CC-Namensnennung - keine kommerzielle Nutzung - 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 10.5446/21220
Herausgeber EuroPython
Erscheinungsjahr 2016
Sprache Englisch

Inhaltliche Metadaten

Fachgebiet Informatik
Abstract Scott Triglia - Protect your users with Circuit Breakers Failures are the bane of scaling a modern web service and can cause serious pain for end users! Lucky for us, there are techniques that can help protect your product and handle failures in subsystems gracefully. This talk will dive into one of these in depth, the Circuit Breaker pattern, and explore the options it gives us for keeping all our users safe. We will be focusing on several real-world problems and options for how to implement your circuit breaker setup in nice, readable python code. ----- The inevitability of failures is the bane of scaling any modern web service and can cause serious pain for end users! Lucky for us, there are techniques that can help protect your product handle failures in subsystems gracefully. This talk will dive into one of these in depth, the Circuit Breaker pattern, and explore the options it gives us for keeping our users safe. We will be focusing on several real-world problems and how they can be addressed by circuit breakers. You should expect to leave the talk with details on simple circuit breakers as well as understanding how they can be adapted for more complex situations. We’ll also discuss some options for how to implement your circuit breaker in readable python.

Ähnliche Filme

Loading...