Agency Workflows

agency workflows only panthéon my name

is David Neum I'm a developer advocate

at Pantheon there's all sorts of ways

you can reach me but really a developer

advocate is sort of a new title for me

and for our team we were previously

something a little bit more complicated

as far as like what we did but basically

what I love about being a developer

advocate at Pantheon is I get to come to

conferences like this and speak to

developers and site builders and people

working with websites and I get to share

with them a little bit about what I know

and I get to kind of help inspire

educate and teach them what I can and

then ultimately if I'm doing my job

right it means that developers are more

successful doing things done more

efficiently and ultimately enjoying what

they do more and so it's kind of a

reciprocal just good energy thing going

on which makes me happy so if you look

at traditional hosting these are really

the three big pillars that you might be

familiar with so sure knows me

you probably already know shared hosting

but really sorry I don't know sorry just

a second let me adjust my screen

okay but so the thing about shared

hosting is it's really cheap and some

clients really cheap is all they got and

that's okay but the problem with shared

hosting is if anyone is having a bad day

everyone on the shared hosting is having

a bad day and unfortunately the opposite

is also true if anyone's having a really

great day and they go viral it also is

really bad for everyone else in the

shared hosting so that's why agencies

and clients often look to something like

a single VM where they can spin up a

server that can kind of manage the whole

thing themself and then they can control

how many different sites are actually on

that virtual machine that's pretty good

it's certainly better than shared

hosting but the problems that you still

end up with a single point of failure if

any one of the things goes wrong if the

database or the Apache configuration or

who knows you push code and it breaks

something then your entire site still

goes down with a single VM and then

scalability also can become a little bit

of a hassle there so that's why bigger

agencies bigger clients often rely on a

cluster which has several application

containers you have you know database

with a redundant database and you have

all this in you know infrastructure set

up which is pretty good I mean it

certainly is like the best option from

these three that you have here but at

the end of the day you kind of have to

have a dedicated DevOps person who is

familiar with architecture and knows how

to set up this configuration and who can

really stay full-time focused on making

sure that security is up-to-date

that everyone has access that needs

access to the right pieces and

ultimately the hardware also it becomes

quite expensive so this is a little bit

of the problem that the founders of

Pantheon originally had they both are

all of them came from agency backgrounds

before and they found that as they were

serving their clients there were all

sorts of different options available and

they all have different preferences they

you know had different needs and that

required different processes to set up

and different workflows and

unfortunately it meant that it was a lot

of overhead for the client so

a lot of expertise to consult and make

sure that everything's set up properly

but also a lot of added expense making

sure that it's the right solution for it

they're looking for and then getting the

developer team set up and ready to go on

Pantheon was also I'm sorry now I'm

talking about on whatever shared hosting

it was how it can become complicated

because you have a different workflow

and a different process for every single

site that you go with so they said okay

there's got to be a better way there's

got to be a better way for us to do this

so they came up with a single

containerized platform when they founded

Pantheon and they said regardless of

what size of site that is that you're

working with whether it's a tiny

mom-and-pop shop or like a personal blog

or whatever all the way up to a huge

enterprise tear web site with thousands

tens of thousands hundreds of thousands

of pageviews doesn't matter we're going

to use the same workflow for agencies

and we're going to use you know the same

sort of like platform where the

experience that you're working with is

going to be the same and then when you

upgrade from one plan to the next it's

just like adding additional containers

but all this is happening under the hood

so if we take a look under the hood and

our actual architecture we are built on

containers we have over a million it's

probably even a lot more about now over

six billion page views a month so again

from the tiny little site a huge huge

enterprise site the architecture is set

up in a way that you don't have to think

so much about what's happening under the

hood we just sign up for an account

focus on the site that you're building

and all of the server performance and

security things are handled

automatically so we're gonna start by

talking about the edge in this case the

edge is what most of the people viewing

your website are actually going to be

viewing this is where the cached version

of your website lives

so with Pantheon we use varnish for our

cache at this level and because of our

new global CDN these are all over the

world so if you are in Berlin and you go

to a web site to toast on Cathy on if

it's cached you'll be viewing the cached

version from our data center from our

our point of presence that's in Germany

and we have

a couple different ones there so you

don't have to hop across the ocean and

back to negotiate that so you end up

with a much faster and more reliable

website but if for some reason the

website is not cached or if you need to

start letting people log in or checking

out the card or anything like that then

we start routing to the runtime matrix

which is made up of all the containers

that make up your site so this is where

your Drupal site actually lives this is

where PHP and engine Nick nginx lives

and then for all of these machines

regardless of what tirion we also have

redundant containers set up so that if

we detect that something isn't quite

right on one of the containers that's

serving up traffic we're able to

seamlessly under the hood without any

sort of interruption whatsoever Rock

traffic to the new duplicate container

spin down the old container and then can

I see what happened

this also means that we can apply

security updates at the server level

without interrupting any sort of any

sites without you downtown or anything

like that so I under that connected to

the runtime matrix we also have marina

DB which is our open source database

solutions just like my sequel we have it

optimized for performance and depending

on the level that you're on we also have

the redundancy in place we have Redis

which is object cache object cache is

ideal for logged in traffic so if you

are if you are logged in you're going to

be hitting the edge if you're not logged

I'm sorry if you're not logged in you're

me hitting the edge and that's the

varnish cache if you're logged in you

can be routed to the runtime matrix and

if you're using reddit

it means that you're logged in traffic

it's also going to be faster because the

database queries are able to be cache as

well and then we have a patchy solar

which if you have needs for search on

your site if it needs to be fast and

full of functionality

basically if search is a key part of

your website you probably want to

consider looking at Apache Solr as an

option we have that built in with our

the business and professional tears so

you just enable it and yeah it's all

kind of the pieces get connected

together and then it's easy to just set

up and then the hallo hallo sort of our

secret sauce so we are able to keep

files in sync across each of your

environments across all of the

containers because of Kahala so even if

you have a website that's several gigs

large you could easily click a button

and clone from your live site down to

your dev site or you know some other

site in just a minute or a few seconds

in some cases and it doesn't matter that

it was several gigs of data and these

things just work so ultimately this this

is what Pantheon consists of this is the

website management platform that is

Pantheon so we end up focusing on the

sorts of things that your clients kind

of take for granted I think often you

know that you have a fast and performant

website that's reliable but you know the

other part of that is also the tools we

provide to developers already talked a

little bit about the tools today when we

look at the different agency workflows

so let's just jump on into that and so

the the this is the basic idea the the

sort of developer workflow that I was

taught even before I came to work for

Pantheon many years before Pantheon was

even a thing for if you're building a

web application having you know at least

a dev site and a live site you know you

don't want to work directly on the live

site because if you make some sort of a

change or if you apply an update and it

doesn't go how you planned they need to

sprint the live website so that's that's

relatively obvious but we also recommend

having a test environment and in this

case the test environment is identical

to the live environment so when you work

on your new feature or a higher security

update or whatever you're doing in dev

you're able to commit that deploy that

up to test and then either manually QA

you know click through the site and test

it or you can even set up automated

updates all right it's our automated

tests that check to make sure that your

website is not going to have any

problems or go down or

unexpected changes when you apply your

updates so all of your commits all of

your code your updates new functionality

is going to start and in dev go up to

test then go up to live on the other

side of that we have syncing content so

we're referring to content in this case

as your database in your files and your

files are like the your files that are

associated with content so you create a

blog post you have an image that you

upload to that or a PDF or something

like that that is a file so your code

goes up from dev test live your content

so your database in your files can come

down and in Pantheon you just click a

button and it syncs up the environment

from one side of the other all right so

let's take a quick look at the Pantheon

dashboard

so when we log in the Pantheon you have

a dashboard that looks just like this

you're able to kind of click through and

see all the sites that you have you can

also see a list of all of your

organizations that you're involved with

so if you are part of an agency you'll

you'll see that there I'm going to skim

through this pretty quickly because

again I want to focus on of having some

time at the end for for questions and

digging into more things later but let

me skip to and just click on a site here

and jump into what that looks like so

here we have a website that I created is

called Needham bikes if I click on View

development site I can actually open it

up and I can see it's the development

sent because it starts with dev - the

name of I site and then dot panthéon

site dot IO so this is what it looks

like in the dev site looks pretty good

if I jump back to my dashboard so I can

see here at the top I have the three

environments out-of-the-box dev test and

live it's easy to click on any one of

these and to click on open the

environment and kind of see what that

environment looks like all right if I

jump over here to the top I also have a

section for workflows let me see if I

can make this a little bigger so I was a

senior for workflows where I can see all

the things that people have done in this

site over time so you know if I looks

like here I created a multi diver

deleted a multi dev environment if I

clear the cache from the dashboard if I

any of these operations their workflows

that happen I'll see when it's happening

with a little spinning ball over here

and I'll be able to click and actually

see who did it and how long ago started

I also have a section here for team

where I can add people to this

particular site it's also possible from

the settings section here to actually

invite someone to pay for your site so

if you are not the person who pays for

the site and whether it's your manager

or the client you're able to come in

here and invite that person to the site

and then they're made the owner they can

select what plan is appropriate and then

or you can suggest what plan and then

they'll put in their information and

actually

this section right here is where you

would also enable solar and Redis and

then you can get some information about

the site including the PHP version the

site is currently running alright so

from here we also have the section where

you can select which development mode

that you're currently working in so in

this case we are in git mode

there's also SFTP mode so if your

developers prefer to work from SFTP if

you prefer to work with tools that kind

of leverage SFTP that's totally fine

there's nothing wrong with that I think

typically SFTP is carried a little bit

of a stigma with it because I think a

lot of people associate SFTP with cowboy

coding you know logging in directly to

the live site and making code changes

directly there but the way the panthan

gonna set up the whole process the whole

workflow is based in get so if you're an

SFTP mode I'm switching now but it's not

important when you're in SFTP mode

you're able to make changes directly to

the server whether it's blogging about

SFTP or using the Drupal interface to

you know upload or update your modules

or you can also connect via drush to

update files directly on the server and

in most cases you're changing files

directly on the server and you'll see

when you make changes that there is

actually a section it will say you know

you have up code changes in this

environment right a commit message a new

clipped commit and it makes you go

through the process of saving a commit

so you can see exactly what's happened

and who's been doing so even if even if

your team isn't all on board with git if

they're if it seems too complicated or

if they're out into that that's okay you

can use SFTP and you still get the

benefit of a commit log you can see who

did each command when it happened and

then whatever the commit message

actually is so it's easy to switch back

and forth between SFTP and get mode and

connect to either one of those that you

want

so for each environment we also have

sections here so we're currently in code

where we can review all the code that

we've made all the changes we've made to

this environment we've also have a

section over for status or we can click

and especially for your live environment

but for any environment you can actually

jump in and see kind of a status report

of what is

going on with this particular site so

it's great to check this to make sure

that you don't have any security updates

that your performance to set up your

caching is set up properly in this case

it's telling me hey I haven't made a

backup recently I might want to consider

doing that we also have a section down

here for database and files so if you

want to maybe copy this new content

that's happened in the live environment

and pull it down into my dev environment

I just have to go over here to say clone

database and files from live I can make

sure those two are checked and then have

the option of choosing either run update

PHP which is good if I'm doing updates

or clear my caches and I just have to

click the button and so now if there

were people adding new content new blog

posts new things to the store on the

live site these are now getting copied

down into my dev site I also have a

section here for import and export if I

would like to do that which can be

really handy if you're migrating or

building the site on Pantheon but then

taking it somewhere else

you can do that through this interface

here we have a section for PHP errors if

you have any of those we can add domains

you can of course create backups which

happens automatically on schedule if

you're on a paid time we'll have a

section here for security where you can

put an additional level of

authentication on top of your site if

you choose to do that so every site that

you spin up with Pantheon comes with New

Relic APM pro out of the box which is a

huge performance tool that that gives

you great insight into why your website

might be running slow it gives you

notifications if it detects that your

website is running slower than you might

expect

and it gives you a really deep look into

like what functions are being called

that cause your website to be slow or

what database queries are being called

all right I mean this is the sidebar

that you see here is pretty much the

same for each of the environments so

whether you're in dev test or live

they're all right there so let me just

call it the elephant the room or there

isn't update available for Drupal and

these updates do appear directly in your

interface here so whenever there's a

core update to Drupal it pops up saying

just like this one says here it says

there's one pops up screen update

available so if I like to apply this all

I have to do is check the boxes that are

appropriate

click apply updates and then it'll get

pulled in I'm not actually gonna do this

right now though because I want to show

you a different way later on in

presentation does anyone have any

questions about this so far okay keep

going

all right so

we go so they're really in this case

we're sort of identifying 10 different

workflows that agencies identify as

being really important to the work that

they do we're going to go through each

of these one by one and we're going to

talk a little bit about what the problem

is what some solutions could be and I'm

going to invite other people to

participate too even though this

presentation is a sponsored session by

Pantheon and we were of course have

solutions to each of these workflows I'm

going to be honest and just tell you

like we're not the only solution there

are lots of other options out there so

if I think one wants to share anything

as we go along please do share all right

first things first agencies want a

efficient set up of tools for new

projects they want to be to get started

on a project really quickly without

having to do a lot of manual work so

Pantheon solution for that is the

developer dashboard so in this case a

whether you're the website that you're

working on is again really really small

or really huge you get the same

developer dashboard that I just demoed a

moment ago you still get dev test live

that are all in place and your

developers can have the same process to

build any site whether it's small or big

or complicated or really easy so one of

our one of our perhaps clients here is

kind of made a quote here Alex a few

more quotes as we go along B uses a

Steve Blackburn of Texas axes here so I

needed a web environment where I didn't

have to manage physical or virtual

servers so I can personally relate to

this as well because I worked for an

agency before Pantheon and in my

situation I was sort of the DevOps guy

by default I did not like DevOps I did

not have any experience really in DevOps

but I knew the most out of anyone on the

dev team therefore I was the person

responsible for making sure everyone had

their workflows in place and had the

server's configured the right way and

invited everyone to the right sites when

we as a company we're looking at

Pantheon as an option

it significantly reduced the work that I

had on my plate basically the process to

spin up a new site for a client we we

got a we want to bid we are about to

start this project just go into the

interface and I say create a new site

and then the workflow that my my team

used for the previous site is the same

as what we're gonna use for this site

it's just consistent and it just works

does anyone have any feedback or

suggestions or any anything that they'd

like to contribute I validate that's the

experience cool good okay well let's

keep on going it's in the next one

agencies really want to be able to share

code between similar sites so a common

use case here is let's say that you are

a university where you have lots and

lots of really similar science whether

it's for different departments or

different student groups or different

classes who knows the typical solution

for this what we like here is something

like multi des I'm sorry not multi

nevikov multi-site like triple multi

site you can share a code base so you

have all the same modules and all this

in some cases the same database you can

break that out and then you put all of

the sites on that one Drupal

installation I've heard I I've only had

the pleasure of working with multi sites

a couple times in my career but I've

heard so many horror stories about it

about if you you need to spin off a site

make it just a little bit different it

becomes really difficult because they're

they're sharing the codebase and then if

anyone wants ID goes down it might take

all of them down or it's all for

different things can anyone relate to

that or have you anyone want to share a

story about working with a multi-site

not without invoking PTSD okay fair

enough

so for Theon we also identified we out

of the box said no multi-site like we

don't support it at all but just

recently your two years ago or so we

released this new feature called custom

upstreams

a custom upstream in this case lets you

have the best of both worlds where you

can share a common code base but you

have the flexibility of having each site

being independently run with its own

containers and with the ability to Nick

changes and customizations however you

want installed profile it can be even a

little more than that so so the way this

starts is you have a the core CMS and

whatever common code you want them to

share just like you might in a custom

upstream and you put this into a github

repo or bitbucket repo or or something

like that and then when you go to create

a site based on this upstream it's just

the same process that you would do in

the pant and face you say create site as

an agency any custom upstreams also

appear in that list so you could say you

know my my fancy edu custom upstream you

click it and it starts the same way that

any other site started so whatever

pre-configured stuff you did you could

even do additional like database changes

and you add this content or whatever

whatever you want to do after that

process and then each site has its own

dedicated site I mean each one is is

isolated from the other so if you have

changes you want to make to it you can

make those changes

maybe net-based right it doesn't I'm

assuming it doesn't automatically blow

up so so so you are you are maintaining

the ever posit ory with the code in it

when an update comes out you can apply

the update to that upstream or to that

that codebase and then when you say that

we need to apply it again then shows up

in the dashboard just like the stream

the pantheon Drupal core update that I

saw earlier it shows up in everyone's

dashboard and they can click update and

so you can distribute updates not just

for core but also any of your contribu

using or themes or any custom stuff that

you're doing you can distribute them out

that way yeah and then it does give you

also the flexibility to make any sort of

custom changes you want so they all

start from the common place and you can

push out changes to the kind of light

from the common starting point but you

still do have the flexibility to say

this one website it started with our

agencies sort of secret sauce and our

theme and stuff but it has a bunch of

extra modules has a it's going in a very

different direction not a problem

because of the way that it's kind of

shipped out you could have a site that's

that's pretty different than the

original and then it also puts you in a

position where if something bad happens

on any one of those websites

it doesn't take down all the websites so

if one website for whatever reason you

maybe maybe the client leaves you know

they fire you and they go off and do

their own thing you're able to you know

give them access to that site you know

kind of put them back on their own path

and it's not a big deal or if they add

their own development team and they make

a change you know they take down their

site it doesn't do anything to the other

sites so you have this

kind of assurance that you know it's not

gonna nothing Bad's gonna happen even if

you have hundreds or even thousands of

different websites running on this click

so cook here from Daniel from hands view

rather than helping people keep their

sites running we'll be able to

concentrate on helping them make

effective websites

so the real takeaway here from this

quote is Daniel is able to focus on

actually building the functionality of

the website instead of having to manage

and the actual things running under the

hood and making sure that all these

websites and a multi site are

functioning the way they were supposed

to okay let's look at the next workflow

developing and deploying different

branches in their own environments this

is a pretty common workflow if your

agency is working with the sort of

typical git flow process the the

thinking here is that you're working on

a project and you have a developer and a

front-end developer and who does a

content creator or something like that

each person should be able to spin up a

branch and work that independently

without worrying about stepping on each

other's toes so you could have a branch

add your custom code another developer

or a front-end developer whatever could

be having their own branch working on

the theme and it doesn't matter so much

that you know they might be using

different you know methods and different

requirements or anything like that when

the time comes for them to be done they

can push that merge it back into the

master branch and it makes it so you can

actually launch with those changes

without without that complexity for our

solution here is multi dev with multi

dev it gives you the ability to have as

many development environments as you

want basically you can go into the

panthéon dashboard you can go into the

multi dev section and you can actually

say create a new multi dev and you could

say you know some common use cases for

that is you know let's give each

developer their own sandbox or let's

create multi dev environment for each

new feature we're working on

it's also pretty common to spin up a

multi-ton environment for security

updates for example and the the takeaway

here the most important piece to this is

that when you're using multi des when

you're using branching it means that

you're able to keep dev tests and live

clear so that if something happens out

of order so for example if a feature

that one developer is working on

finishes before the other one or if a

security update comes out in the middle

of a development process you're able to

apply that update you know create a new

one apply the update apply it to dev and

not worry about oh is this code that

we're working on ready to go to live is

it is it actually ready like can we do

we have to roll it back more likes to

rollback code no one likes to undo their

work less that's no fun so keeping dev

test and live clear and available for

updates means if John Doe my coworker

you know finishes this feature before I

finish my theme he can apply that test

it and deploy it without having to wait

on me

so patch comm is a client of Pantheon's

and this is a pretty cool quote because

he mentions you know 50 million

pageviews 20 million uniques at that

uptime with 35 million URLs and 40

terabytes of images all with four

developers it's really speaks to the

sort of power and scalability that

Pantheon lets developers have it doesn't

matter if you have lots lots of

developers let me switch back to the

demo so in my dashboard this is the site

that I'm working on let me show you a

multi dev workflow so in this example

here I have well I have a security

update I just upstream update available

the way that I recommend applying it is

through multi dev environment so let me

show you how that works

so over here at the top next to dev I

have a section of this multi dev if I

click on that I'll see all of the dev

environments or I'm sorry multi dev

environments that I currently have so I

haven't we're called core updates and it

currently says hey this is one behind so

that means one commit has been deployed

to the dev environment since the last

time I've synced up so it's it's been

behind not a big deal it's zero ahead

which means there's no commits in this

environment that haven't been applied to

debía so right let me go to click on

core updates and I can see right so hey

there's one commit that can emerge from

master and I can see exactly what that

update was so it was actually a previous

version of Drupal that was updated and

so that's that's cool no problem I'm

gonna start just by merging this update

that's already been applied into this

environment so that our multi dev

environment is actually going to match

master and then what I'm going to do is

I'm going to apply this upstream update

the core update under this environment

so I can test it make sure that I know

what's going to happen my actually merge

it into

the Devon burger so I'll start just by

clicking merge at this point it is it's

really kind of just running get commands

under the hood so if you're familiar

with git that's that's all this this

process is doing it's looking at dev

which is master it's looking at the core

updates branch and it's yeah putting the

master on top so yeah so so there is

this option here that says auto resolve

conflicts um so in a case where it's a

security update like this you you

already probably know that you shouldn't

be hacking for so this would if you did

have core it would overwrite that in a

case where it's just like the merge like

I just did if you had you were working a

feature your colleague is working a

feature and you change the same file

that's complicated to do for my UI

sometimes you you there may not be a way

to do it that way it doesn't happen as

often as I expected it to when I first

started but when it does happen you kind

of have to either basically the git pull

you know pull it locally resolve the

conflict and push it back that's that's

really the easiest way but so it's not

finished so I have now this this

environment should be just like my dev

environment I can actually click this

this link here to open up my custom

upstream environment is to make sure

that says core updates so that's the

environment name and then the name of my

sight and then Campion side to i/o so

everything looks good just like I expect

it to so now I'm going to go ahead and

apply this I'm gonna check the box for

all the resolve updates even though I

don't expect it would make any

difference one way or the other I'm not

going to apply the updates so in this

case this is just a Drupal core update

but you can imagine the scenario where

you have module updates where you know

every Wednesday some new updates are

released this is the process you might

go through when you're applying those to

put them into a multi dev a very safe

space to work and you can then do

whatever tests you need to do you can

even talk a little bit later about

scripts you can set up to automate

certain tests

you could say whatever an update comes

out applied to a multi dev do a visual

regression test to compare screenshots

from this site versus the live site

if there are no discrepancies everything

is exactly the same let's go ahead and

deploy that up to the live site because

we know nothing nothing broke all right

all right so it says here now that that

message went away because we applied

that update I have one unmerged commit

which is this update 28.5 and so this is

the player at which I would go through

and I'd run my QA I would first check to

make sure on the corbs inside I'm this

case I'm manually just kind of checking

because I feel pretty safe about it but

you might have a whole process to make

sure that the updates you are applying

aren't actually breaking anything I feel

pretty good about this I'm confident

that this is going to work so with that

in mind I'm ready to merge this in I can

either go back to the multitude of

overview where it says hey now you're

one ahead click here to merge into

master or I can go to the dev site that

can merge from there I'm just going to

do it from here and save a step so it's

asking me what Monty doesn't want to

merge from core updates and click merge

code

but so well this is really let me just

demonstrate another thing so I have

another multi dev set up here that this

feature one which is three behind and

like other stuff is going on you can

imagine where there was a long-running

feature where there's lots lots of

updates well I just added a new

developer to the team and they have a

local environment set up and they spun

up a branch and they did some work and

then they pushed it push that branch

into Pantheon and so using the section

here on the side I can see all the get

branches that people have pushed up to

Pantheon and then with a click I can

create an environment based on this new

branch that was just pushed here well so

at any point

anyone could create a new a new branch

with their code and their custom stuff

push it up and then I can clone a

database from whatever environment and

get it up and running and then they'll

have a dedicated place to do

configuration or to demo to the client

or to get code review from their

colleagues and I can see from their

workflows spinning here that this is

running and there's stuff happening it

might still be merging as well the merge

finished and I can see I have the update

right there and so the next product

process of this the next step would be

we have these changes in dev done we

need to apply them build test and live

on see we just go to test threads

deployment message in this case it's

Drupal to Drupal security updates

and I have options here to auto-sync up

the content from a live site down to the

test site in this case and I'm not too

concerned about doing that so we're

going to run up to get update dot PHP

through the cache when I do this click

the button

so this is going that well this is going

I'm gonna go ahead it just takes a

minute I'm gonna switch back to the

presentation so in the interest of time

let's just keep on going and I'll come

back to that in a minute all right

the next workflow is optimizing the

infrastructure for performance also the

good news here is that we kind of

already just do this you don't have to

think too much about tuning performance

the the most significant performance

boost you can give to your site is

actually just enabling caching in the

interface once that's done you're

talking about like millisecond or sub

second performance tuning there are

things you can do to get into that and

we certainly have recommendations on our

website and entire class is dedicated to

doing that but a lot of the things are

just already set up by default you don't

think about it you just decide another

quote about going the interest of time

the next one adding and removing team

members from a project also this is

something like I mentioned before was a

kind of pain for me coming from I have

my own server I need to go in and add a

new SH key for the client or for that

the developer I'm working with and if

they're a contractor I need to make sure

that you know kind of give them the

tight reins I mean maybe remove someone

who just left or all those things and I

need to do that for every project we're

working on which ends up even with a lot

of time so for Pantheon's interface we

have the team management console by

which you can get from the agency

dashboard so let me switch back to that

here so this is where the agency

dashboard looks like you see all the

sites that this agency is responsible

for listed here and we see all the

different filters we can use from here

we also have a section for people so any

people that you add into this list will

have access to the sites that we saw in

the previous page by default but we also

have options here when adding someone to

actually change their role so you could

say on some limited developer as a

developer you're able to work in a that

multi dev environment but you're not

actually able to deploy to your tasks or

live site so this is the case where you

know you're okay with someone working on

some new features that maybe needs to go

through some sort of approval process

you want a senior developer or the

project manager to a sign off on these

changes first the team member has the

ability to do them they can also manage

the site level team which means for each

site they can independently assign

particular people so for example if you

hire a contractor they don't need access

to all the sites they just need access

to like one site you can go into that

particular site give them access set

their role and then they're good to go

well then the administrator of the

agency can do a lot more things with the

agency itself like change the name or

add more people

alright and they have another quote here

for being residents I went to my project

manager and asked if she had scheduled

up time to start building the singing

boxes for a project she said she had

already taken care of it herself that

was the project manager had already

taken care of it herself that was the

moment of whoa this is real it was my

big moment I can certainly relate to

that the next workflow actually

launching the site and then all the

subsequent you know iterations and

deployments that you make so that for

that we have Deb test live as well as of

course the multi devs that we talked

about earlier we do depending on the

tier if you have like an enterprise

level site or we also have a guaranteed

launch process where we have some

dedicated to helping you make sure that

your website is going to go smoothly and

is on the right tier and has performance

testing all the things set up in place

so that's not a problem

he's working from another quote from

white fused media here so I can actually

relate to this again from the agency

side because before when I was working

for the agency the process to take a

website live it would often take weeks

of scheduling because you'd have to make

sure you know talking to the client

making sure that the whatever solution

they chose or whatever solution that

you're recommending the servers set up

properly that is tuned for Drupal that

it is going to work with all the

functionality that the site actually is

supposed to have and then you actually

have to log in and test it and make sure

and then you have to configure and make

sure when it when the traffic hits that

it's not going to take the site there's

all these steps you have to do

which means you're spending a lot of

extra time making sure that the websites

going to launch properly when the agency

I work for switch to Pantheon we went

from it was it was a noticeably stark

difference because we could say yeah go

check out the live URL this is exactly

what's going to look like we can

performance test that live URL

it's good to go when you want to go live

just change the DNS settings and it's

it's gonna go there's no stress there's

no worry just do it once done

all right so applying court dates I

actually already talked about this there

is that built-in one-click or updates

process and again this is something it

could be automated if you so choose to

keep on moving

connecting continuous integration

processes so this is a pretty cool thing

we're actually teaching a class at

Drupal con Nashville that will go

through the process of working with a

Drupal 8 site built in composer hosted

are not hosted but with a code base on

github built with circle see I deployed

to Pantheon so that all sounds kind of

complicated but built into that process

is the dedicated build step with a tool

like continuous like a circle CI or

Travis or something like that and then

using tools like our command line

interface with terminus we're able to

connect from that environment over to

Pantheon we're able to run b-hat tests

or other tests that are important to us

on that environment and then automate a

lot of these a lot of these things that

are happening under the hood

alright and then one of these last moves

here is actually notify external systems

about deployment progress and for this

we have a tool called Quicksilver

platform hooks with Quicksilver you're

able to basically inject your PHP code

between different workflows in the

panthéon dashboard so whenever you make

a deployment whenever you make a commit

whenever you clone the database in the

panthéon interface you're able to inject

this code in there to communicate to

these third-party systems so you could

say every time that I deployed alive

post a message in slack you know letting

the client know every time that we clone

the database the live database let's

scrub the database of private

information like if it's an e-commerce

site you want to get

there's credit cards or anything like

that client information you could do

that

every time that I make commits with a

particular hashtag number format

automatically post a message in JIRA

with a link to the multi dev letting

someone know that there was work done on

that task other the options are trying

to limit lists we have a huge repository

of examples that you can dig into and

actually take a look at and kind of see

and use for your own this is an example

of another one so for this particular

this isn't looking at New Relic so it's

again performance monitoring based on

this every line in this graph is it

commits for a deployment that we've done

and so you can see basically where I

don't know exactly the scenario spent

remember I set this up but it looks like

there may be deploying a change and then

maybe undoing the change and then doing

it's kind of a complicated thing but

basically you can it communicates the

New Relic so you can see exactly when a

commit happens and then what the result

is what happens after this commit

changes so you why is my website running

slow you go into New Relic you have are

showing hey there's a commit five

minutes ago who did this

what did they do what did we change

about that and it gives you another

place to look to when you're debugging

that

okay reorient in a new project this is

also kind of a no-brainer because it

just kind of works it's once you've

built a website on Pantheon and you're

use that workflow it's the same process

regardless of the size or complexity so

it just keeps going alright so we have

another quote here and let me just recap

here these are the ten that we just

stepped through before we wrap it up let

me switch back to my example so I have

here our test it's been deployed to test

if I go into live I can see oh there's a

commit ready to deploy from the live

environment but before I do that

remember I said test is we were supposed

to really do these the the checking to

make sure that you know when you deploy

it nothing nothing's going to happen

that's unexpected so I actually set up

some automated tests in in the test

environment it's set up with Quicksilver

so that every time I deployed a test it

automatically goes out to a third party

like load testing website it hits the

test environment with a bunch of traffic

and it will tell me how it responds to

that traffic so if I open up its website

called loader IO I can see here on March

is probably when the deploy completed

and I can see what happened so it looks

like everything is probably pretty

normal like look I've I've done this

several times this is a pretty normal

result and so like I you know without

having to do anything extra this just

happened for me and you could do

additional steps to say you know every

time that you deploy run a test and then

put the results in slack so if you've

been to Drupal con if you've seen our

booth demo that's one of the things we

really like to showcase because it's

something we're really proud of it looks

really impressive but then also if we go

into New Relic

so there's co-pays single sign-on bug

that we have to work through but the way

it's supposed to work is you click on

this go to new relic button it takes you

into new relic and then you can actually

see all the environments that you've set

up so based on this load test I should

be able to see performance data

specifically for that environment and

then see how it works

assuming everything passed which I think

it did we can go into live and actually

deploy the change live

click the play and then we're good to go

all right with that that's they do it

driving questions further advanced to

automate configuration so the question

was are there any plans to automate

configuration management um the good

news is that you already can so you you

set up the code so using Quicksilver the

automation hooks hooks feature you can

basically set it up so that every time

that you commit or deploy it will

automatically run rush and do a import

of the configuration and then you just

deploy it like you normally would and it

automatically imports any configuration

that you've exported okay so so you

would have to use the command line every

time you just set it up and then it's

good to go

yes it could you recap when you would

want to use alternate workflows than

what has survived you fall sure so I

mean the the question was can I recap

any alternative workflows that might be

different from what we talked about here

so so one of them that's freshest on my

mind is the kind of like the training

that we're doing in in Drupal con where

our code base is listed on github and

the process involves I'm working just

with github basically I'm committing I'm

pushing everything to get hub and then

based on the few that I make there it

auto deploys that to Pantheon and then I

so that your developers they never

actually log into Pantheon it's solid

say they're not to do with the code

there at all so that's a very unusual

workflow from from the from the norm but

the benefit is that if you have a

process where you're working with a lot

of developers and you want a

standardized build step circle CI is

definitely the way to or a continuous

integration process gives you that and

then that gives you a great tool like

when you make a commit and you you make

a bunch of commits and then you make a

pull request so it could actually notify

your senior developer and say hey

there's this pull request you know

through github like you normally would

but then all the tests will be right

there with a check mark or an X to say

did the code standardization you know

past it the visual regression test pass

did the B hat test pass and then I'll

give you a link directly to the multi

dev so they can go and actually look at

it so gives you give them a great

interface to say like yes this is good

or no I need to you know push it back or

something all right so the short side of

that is in a way you could have

basically

a github based workflow going on for say

is made out the seniors they approve it

or not but once they accept it then

antion's

hooks can approve it and then working

into the deployment side right right yep

yep so so as they're making these

commits they actually multi dev

environments might be spun on so they

can actually review their work or make

configuration or whatever but then when

you do a pull request and that's

accepted it actually detects the

branches that were created destroys

those you know merges it into the the

dev environment as well yeah yeah and it

also can circle back the other direction

so when it's like let's say you make a

command it spins if a multi dad if you

go into that multi dev and then export

your configuration that can get pushed

back to github and show up as a commit

as well

so time you can really go any way you

want

that's just one of the more popular

alternatives that we've been seeing I

admit it's like in the passive fruit

that pantry they explode on what the it

was and so I was kind of wondering if

that was still okay sir

all right cool thank you all very much