DrupalCon Portland 2013

Lo que debes y no debes hacer al testear el rendimiento de tus aplicaciones

Erik Webb, Jeff Beeman  · 


PDF (pincha para descargar)



Extracto de la transcripción automática del vídeo realizada por YouTube.

so and I'm Joe Keenan I'm also a senior technical consulting with back via my focus not performance ability on a day-to-day basis and sort of the reason I'm here is to talk about why that really mattered because I have to care about performance

anyways I found out via in 2010 so a little bit after Eric I've been working with triple for quite a long time and my primary role on projects is usually as so in that role actually how many you guys are you say but I think that's actually the team's

rising really so before we get too far in the discussion it's really important what's the level set on some terminology and talk about the various things that go into performance testing because like Eric was kind of hinting at performance testing

isn't just looked at it's not just loading a page on your site and seeing how fast it is like that there there are lots of aspects so we're going to talk about some terminology and some of it you make you might use their words to talk about these

things but Google cleared lines in your minds about what's moving Sinclair performance is about the speed or the time it takes to do an individual request this can be both from the time a user requests a page to the amount of time that it takes for that

piece of the map but there are also other things that we're going to measuring performance engage a long did it take to render long does it take for all the assets are there javascript sexting on the page that make it take a long time so you actually fully

rendered less performance scalability is the ability of your site to go from one user to 100 users or 100 users to a thousand users and still have a predictable performance and predictable being like not the same 100 users hitting their site is probably gonna

perform a little bit differently than a thousand users in yourself or 10,000 users doing your site but it's acceptable and it's predictable there was a lot of discussion like drupal 7 came out about how Drupal 7 was slower and the whole idea what it

could skate all much further and prior versions of truth so just because an individual page requested flower doesn't mean your site is performing works my perspective give up some speed on individual page across northern middle a enormous volume of time

is scalable so scaling is how it comes across in your community it's probably the area most people start with one that will be advertised scaling is saying we want to take our tutors call servers and we want to make them bigger that's scaling vertically

your scale horizontally but sticking or two small servers down there in the corner and just adding more so load testing is about focusing on real world traffic problems and seeing what happens to your site for measurable amounts so in five here we've got

10 users hitting our site and our servers happy when we throw under users that website servers you know straightening a little bit it's having a little bit of a hard time load testing is not about breaking load testing the same I want to throw eggs number

of users of the site and we're going to measure and see what happens stress testing is about seeing what happens or how far you can push your site before it breaks other people call it's like a smoke test and stress testing is about saying like how

many people can we throw out the site before it crashes and burns I like to think about like a load test being like how much does I'm going to get a scale I'm going to see how much 10 people with and we measure that can save people way but a stress

test is taking a scale and putting people on it until scale based last one that we're going to talk about we're going to talk about a lot that there is a sort of low wattage contention contention is a pipe over resources and there are lots of places

that you can end up adding contention in your site it's we promise to be the hardest thing to like draw a diagram of it's probably the hardest thing to talk about but it's also the most important concept to understand when you're thinking about

I think one way you can think about the convention is to think about walmart on black friday so a whole bunch of people a civ people line up at the doors and then they open this tiny doorway that's about four people across into the store and everybody

wants to get anyone's and it would fight over the resources both to get in and then there's a fight over the research side there are many points of contention you go from one server to servers theoretically militate do the first server just whether

that's for your continued life and then a second circuit doesn't actually help the problem that's why I had soaked even identify where the containers if that is what and we've got something we've got a couple slides a little bit later at

the convention is the concept you don't understand that effect explain it a little bit more but right now just try to think of it as a bite / usual hopefully up your understand so let's talk about what your tips are actually going to do so we kind

of hurt right campus present a little bit to what you should do and also gotta lessen fine but there's meeting we've worked on in the past and what we like the dog bit they succeeded so the key is that you're tense not good the number to get for

them apart is simply a welcome if your test a fantastic that you can use those numbers and actually decide you're going to service how much time you spend on further tests that the quality of the test is really valuable piece make sure you're setting

technical roles so don't look at this as what makes it user trying to guess what writers actually track these things that required look at you add feature eggs what what's actually what are you doing actually monitor that change side that's not

you're not looking at a side by side it's really easy to get 3 quarters of the way to project and really have no idea what that bottle against refund the biggest so far down the line the project unit three more human resource set up to solve problem

really could process so regressions from the QA on the typical q icing so features breaking other features it's just as easy the new features so anything they can thank you out the way of what makes the good test how much you're testing how wide that

[ ... ]

Nota: se han omitido las otras 3.230 palabras de la transcripción completa para cumplir con las normas de «uso razonable» de YouTube.