Global Azure Bootcamp 2017 at Betabit Utrecht

© Betabit Nederland B.V.
Daniël te Winkel
d.te.winkel@betabit.nl

Dagindeling
10:00 Start deel 1
12:00 Lunch
13:00 Start deel 2
16:30 Afsluiting (& prijsuitreiking!)
17:00 Einde Global Azure Bootcamp, locatie Betabit Utrecht

Agenda – Deel 1
- Intro
- Use case
- Web Apps
- Web Jobs
- Hands-on Lab
- Functions
- Hands-on Lab
- Service Fabric

Intro

Even voorstellen
Het team:
• Daniël te Winkel
• Bart Kemps
• Herbert Lindner
• Herman Lindner
• Raymond den Haan

Resiliency

Azure: services will fail
• Meerdere instanties van applicatie
• Gebruik een Load Balancer / Azure Application Gateway
• Gebruik Availability Sets
• Automatisch schalen om aan hogere vraag te kunnen voldoen
• Werk asynchroon waar mogelijk
• Gebruik ‘resiliency patterns’ voor externe communicatie

RTO, RPO & MTTR
• Recovery Time Objective
Acceptabele maximum tijd dat een applicatie niet beschikbaar is na een incident
• Recovery Point Objective
Acceptabele maximum tijd dat data verloren gaat tijdens een incident
• Mean Time To Recover
Gemiddelde tijd die nodig is een applicatie te herstellen na een incident

Azure SLA
SLA Downtime per week Downtime per month Downtime per year
99% 1.68 hours 7.2 hours 3.65 days
99.9% 10.1 minutes 43.2 minutes 8.76 hours
99.95% 5 minutes 21.6 minutes 4.38 hours
99.99% 1.01 minutes 4.32 minutes 52.56 minutes
99.999% 6 seconds 25.9 seconds 5.26 minutes

Composite SLA’s

Use case

Web Apps

Azure App Service Family
Web Apps
Web apps that scale with
your business
Mobile Apps
Build mobile apps for any
device
Logic Apps
Automate business processes
across SaaS and on-premises
API Apps
Build and consume APIs in
the cloud

Azure Web Apps
Support a variety of languages and platforms
.NET, Java, Node.js, PHP, Python, and more
Support scaling (manual or auto) and load balancing
Support slots for staged deployments and A/B testing
Support continuous integration
Global Scale
Scale up and down as needed,
manually or automatically
Enterprise Grade
ISO-, SOC2-, and PCO-compliant
with enterprise-level SLAs
Familiar and Fast
Leverage existing skills, plus
languages, frameworks, and
tools you're familiar with

Scaling - Cloud Computing Patterns

Scaling Up vs. Scaling Out
Scale Up
Vary the VM size
1 Core w/ 1.75 GB RAM
2 Cores w/ 3.5 GB RAM
4 Cores w/ 7 GB RAM
Scale Out
Vary the VM count
Max 3* instances
Max 10 instances
Max 20/50** instances

Manual Scaling vs. Auto-Scaling
Auto – Schedule &
Performance Rules
Manual – Scale via portal
or scripts
Auto – CPU Percentage

Deployment Slots
Use a Deploy-Confirm-Promote workflow
Promote via “swap” through Azure portal
http://sitename-slotname.azurewebsites.net

A/B Testing
Test changes by routing requests to different deployment slots
Use Traffic Routing to direct % of traffic to alternate slots

Continuous Integration
Web apps can be deployed manually via FTP or WebDeploy
Automate deployment using 3rd party source-control providers
Can also use a local Git repository from Azure Portal
DropBoxBitBucketGitHubCodePlexVisual Studio
Team Services
Git

Continuous Integration + Deployment Slots
1. Developer commits code
2. Automated process builds/compiles
and deploys to staging slot
3. Automated and other tests validate
content in staging slot
4. Staging content promoted to
production
5. Users see updated site
Staging
Production
TestSource
Control
1
2
3
4
5

App Service Plans
Billing and provisioning for App Service resources
Free Shared Basic Standard Premium
# of Apps 10 100 Unlimited Unlimited Unlimited
Shared Disk Space 1 GB 1 GB 10 GB 50 GB 500 GB
Maximum Instances 1 1 3 10 50
Autoscale No No No Yes Yes
Staging Environments 5 20
Custom Domains No Yes Yes Yes Yes
SLA 99.95%

Resiliency tips & tricks
Kies de juiste tier en scale out. Scale up kan reset veroorzaken
Geef elke omgeving (OTAP) zijn eigen App Service Plan
Scheidt web apps van web APIs voor afzonderlijke schaling
Gebruik App Service backup niet voor SQL backup (creëert .bacpac, kost DTU’s)
Gebruik Deployment Slots

Web Jobs

Azure WebJobs
“Run programs or scripts in WebJobs in your Azure App Service web app in three ways:
on demand, continuously, or on a schedule.
There is no additional cost to use WebJobs.”

Azure WebJobs
Host and schedule lightweights job on an existing website
Triggered once, by schedule or run continuously
Support:
batch (.exe/.cmd/.bat)
bash (.sh)
javascript (.js as node.js)
php (.php)
python (.py)
Dashboard Experience

Store & start WebJobs
WebJobs are stored within the website
• SitewwwrootApp_Datajobs{job_type}{job_name}
Will find the file to run by (first one found wins):
• First look for a file called run.{supported extension}
• If not found, look for any file with supported extension
• If not found, then not a runnable WebJob

Many Options for Deploying
• Azure Portal
• Visual Studio
• Visual Studio Team Services – Build
• FTP
• Dropbox
• Drag-n-drop in Kudu
• Others …

WebJobs SDK

WebJobs SDK
“The purpose of the WebJobs SDK is to simplify the code you write for common tasks that a
WebJob can perform, such as image processing, queue processing, RSS aggregation, file
maintenance, and sending emails. The WebJobs SDK has built-in features for working with
Azure Storage and Service Bus, for scheduling tasks and handling errors, and for many
other common scenarios.”

WebJobs SDK components
NuGet packages.
NuGet packages that you add to a Visual Studio Console Application project provide a
framework that your code uses by decorating your methods with WebJobs SDK
attributes.
Dashboard.
Part of the WebJobs SDK is included in Azure App Service and provides rich monitoring
and diagnostics for programs that use the NuGet packages. You don't have to write
code to use these monitoring and diagnostics features.

Triggers
Poll queue & create blob for each message:

Scheduling
TimerTrigger can be used to run WebJob on a schedule
Schedule WebJob as a whole (through Azure)
Use TimerTrigger to schedule individual functions of a WebJob

Dashboard features
Invoke functions
Replay functions
View logs
- Linked to WebJob (Console.Out, Trace, etc.)
- Linked to function invocation

Azure Functions

Azure Functions
Create a “serverless” event-
driven experience that
extends the existing Azure
App Service platform by
building “nanoservices” that
can scale based on demand

Supported Languages and Tools
Create functions in JavaScript,
C#, Python, and PHP, as well
as scripting options such as
Bash, Batch, and PowerShell,
that can be triggered by
virtually any event in Azure,
3rd party services, or on
premise systems

Common Scenarios
Your App or
Service
Office 365
Office
Graph
Azure
Storage
Other
Functions
Legacy
Systems
Web
Services
Timer-based processing
Azure service event processing
SaaS event processing
Serverless web application architectures
Serverless mobile backends
Real-time stream processing
Real-time bot messaging

Function App Templates
Function App templates are categorized into
general areas of Timer, Data Processing, and
Webhook & API
• BlobTrigger
• EventHubTrigger
• Generic webhook
• GitHub webhook
• HTTPTrigger
• QueueTrigger
• ServiceBusQueueTrigger
• ServiceBusTopicTrigger
• TimerTrigger
• Blank & Experimental

Timer Function Apps
Run at explicitly specified intervals, like every day at 2:00 am using CRON
expressions, like “0 */5 * * * *“ (every 5 minutes)
Can send information to other systems, but typically don’t “return” information, only
write to logs
Great for redundant cleanup and data management
Great for checking state of services
Can be combined with other functions

Data Processing Function Apps
Run when triggered by a data event, such as an item being added to a queue or
container
Typically have in and out parameters
Great for responding to CRUD events
Great for performing CRUD events
Great for moving content
Access data across services

Webhook & API Function Apps
Triggered by events in other services, like GitHub, Team Foundation Services, Office
365, OneDrive, Microsoft PowerApps
Takes in a request and sends back a response
Often mimic Web API and legacy web services flows
Typically need CORS settings managed
Best for exposing functionality to other apps and services
Great for building Logic Apps

code
Anatomy of a Function
A “Run” file that containing the function code
A “Function” file containing all service and trigger
bindings and parameters
A “Project” file containing project assembly and
NuGet package references
App Service settings, such as connection strings
and API keys
.NET Core and Project
references
Function configuration
Executable code

Function Bindings
Type Service Trigger Input Output
Schedule Azure Functions ✔
HTTP (REST or webhook) Azure Functions ✔ ✔*
Blob Storage Azure Storage ✔ ✔ ✔
Events Azure Event Hubs ✔ ✔
Queues Azure Storage ✔ ✔
Tables Azure Storage ✔ ✔
Tables Azure Mobile Apps ✔ ✔
No-SQL DB Azure DocumentDB ✔ ✔
Push Notifications Azure Notification Hubs ✔
Bindings serve as the basis for all connections to and from a function. Many
bindings can be “bi-directional” as well.

Testing Functions
Command-line tools
3rd party products such as
Postman and Swagger
Direct web calls via cURL
Nested functions
Microsoft Azure Storage Explorer
Visual Studio Cloud Explorer

Hands-on lab

Service Fabric

Topics
Introducing Microservices & Azure Service Fabric
Service Fabric Concepts
Developing Service Fabric Applications
Deployments & Upgrades

Introducing Microservices &
Azure Service Fabric

Microservices Architecture
An evolution of Service Oriented Architecture
Create a system as a suite of small, independently deployable services
Services communicate across lightweight protocols
Considerations
Services can be scaled to meet their individual needs
Supports use of different technology stacks for different system parts
Process automation is important (DevOps)
UI/UX
Storage
Storage
Business
Logic
Business
Logic
1...n
Monolithic Architecture
UI/UX
StorageStorage
Shopping
Cart
AuthN
Microservice Architecture
Storage
Product
Catalog

Azure Service Fabric
Can be thought of as an orchestrator of services across a managed cluster of
machines
Good for microservices, but not only
Can be run in Azure, on-premises, or in other clouds
Proven platform used by several Microsoft Cloud services, including:
• Azure SQL Database
• Document DB
• Cortana
• Microsoft Power BI
• Microsoft Intune
• Azure Event Hubs
• Azure IoT Hub
• Skype for Business

Application
Service Service Service
Node Node Node Node
Cluster
Partitions

Service Models
Reliable Service & Reliable
Actor Services
• Leverage Service Fabric’s
own programming models
• Networking Naming
Service support
• Integrate with code,
configuration, and data
upgrades
• Instances are not process
isolated for each instance,
but instead are created as
objects (higher density on
host.)
Guest Executable Services
• Arbitrary executable
packaged in a Service
Fabric service
• Can be written in any
language
• Service Fabric manages
orchestration and
execution management of
the executable
• Limited integration with
the Service Fabric APIs
Container Services
• Containers can either be
Reliable Services or Guest
Containers
• Supports Docker
Containers (Linux) and
Windows Server
Containers
• Currently in Preview

Reliable Services
Stateless Service
• Typically used for Web API front-
ends or background workers that
look at external queues
• No state is maintained within the
service, though external state
storage can certainly be used
Stateful Service
• Used when state must be consistent
and present for the service to
function
• Leverages the Reliable Collections
• ReliableQueue, ReliableDictionary
• State is kept locally, but replicated
for high-availability, backed with
disk storage for durability, and
transactional.

Partitioning – Stateless Services
Stateless Services
Node 1
Node 3Node 2
Node 5Node 4
I
I I
I I
• In a stateless environment,
scalability and availability generally
achieved by adding instances
• Place an instance of the service in
each Node
• Add Nodes to scale out

Partitioning – Stateful Services
Stateful Services
Node 1
Node 3Node 2
Node 5Node 4
S3P1 S2
S1P3
S4
S3P5 S4
P2
S5
S5
P4S1 S2
• For stateful services
• Services are Distributed for
scalability
• Service state is Replicated for
availability
• Distribute primary replicas across
the nodes in the cluster and also
place secondary replicas
• Add Nodes, then rebalance to scale
out

Reliable Actors
Implementation of the Virtual Actor pattern
Ideal when the application involves many (thousands) of small isolated units of logic & state
Built on top of Stateful Reliable Services
Actors are instantiated upon reference and collected upon inactivity
State Management can be used during reference to “restore” an actor
Actor actions are turn-based (single-threaded) per instance.
Supports Timers and Reminders
Also supports Events

Developing Service Fabric
Applications
Windows Setup

Installation
The Microsoft Azure Service Fabric SDK is included in the “Azure development”
Workload in the Visual Studio 2017 installer

Creating a Project
Create a new project by selecting the “Service Fabric Application” template from either
the Cloud project node.
Choose your desired service type template

Solution Structure
Application Project
The Services node lists the service projects that make up the
Application
Includes an Application Manifest that describes the services that make
up the Application
Includes application publishing profiles, and per-profile settings.
Service Project(s)
Includes a Service Manifest that describes the Service characteristics
Includes the Service code and configuration

Service Fabric Local Cluster
Installed as part of the Service Fabric SDK
Provides a local debug & test environment
Can provision a 1-node or 5-node cluster
Not an emulator or a simulator
Can be managed via the Windows System Tray application or via
PowerShell scripts.
Visual Studio is configured to work with the Local Cluster

Service Fabric Explorer

Deployments & Upgrades

Deploying in Azure
Service Fabric Applications are deployed in Azure into a Virtual Machine Cluster in a
Virtual Machine Scale Set (VMSS)
Clusters can be created on VMs running Windows Server 2012 R2, Windows Server
2016, or Linux Ubuntu 16.04 (in preview)
Clusters networking can be configured via Azure VNets, Subnets, Public IP Addresses,
and internal and/or external load balancing
Can be managed with Azure ARM templates,
Auto-scaling support is provided by Azure VMSS
Management is via a combination of Azure tools and Service Fabric tools

Application Deployments
Two types of deployments in Service Fabric – Full and Upgrade
In a Full Deployment, the entire application is torn down (if already present) and then a
new application instance is deployed.
Upgrade Deployments
Service Fabric supports rolling upgrades to maintain uptime
Service Fabric monitors application health during the upgrade. If the system violates the
application health policy, the upgrade is rolled back.
Several modes govern the automation and health-check behavior during an upgrade.

Lunch

Agenda – Deel 2
- Cognitive Services
- Hands-on lab
- Visual Studio Team Services
- Source control
- Builds
- Release Management

Cognitive Services

Cognitive Services
Enable natural and contextual interaction
Powerful artificial intelligence algorithms
Vision, Language, Speech, Search & Knowledge
Free trial
Subscription management
(products, quota left, add additional products)
Upgrade to unlock limits
Regional availability

Vision

Language

Speech

Search

Knowledge

Computer Vision
Tagging images
Categorizing images
Identify image types
Domain-specific content
Generating descriptions
Perceiving color schemes
Flagging adult content
Optical Character Recognition
Recognize handwritten tekst
Generating thumbnails

The 86-category concept

Computer Vision – Example

Generating description – Example

Face API
Face detection
Face recognition
Face verification
Finding similar face
Face grouping
Face identification
Face storage

Face API – Example
Face detection
Face Grouping

Cognitive Services – How to use…
They are ‘just’ REST APIs
Use SDKs

Visual Studio Team Services

Source control
Team Foundation Version Control (centralized)
Git (distributed)
The most commonly used version control system today
“Git is the default version control provider for new projects. You should use Git for version
control in your projects unless you have a specific need for centralized version control
features in TFVC.”
- visualstudio.com documentation

Continuous integration
- Link builds to test results, work items, code and releases
- Hosted build agents
- Multi-platform (Windows, Mac, Linux)
- Security, pool and queue management
- Automated builds
- Diagnostics
- Continuous deployment

Automated deployments
Design and automate release pipelines
across your environments to any
target platform.

Azure
Setup your release with built
in tasks for Azure and Azure
Stack – deploy to websites,
Docker containers,
virtual machines,
and more.

Quality
Configure tasks for all of your
release check points –
performance, A/B,
functional, security, beta
testing and more.
No more release day nightmares.

Control your deployments
Enable sign-off for
deployments using pre or
post deployment approvals.
Automatic notifications
ensure collaboration and
release visibility among
team members.

Traceability
Track the status of releases and
deployments including commits
and work items in each
environment.

Resources
Azure regions (includes products per region)
Cognitive Services SDKs & Samples

Global Azure Bootcamp 2017 at Betabit Utrecht

Empfohlen

Empfohlen

Weitere ähnliche Inhalte

Was ist angesagt?

Was ist angesagt? (20)

Ähnlich wie Global Azure Bootcamp 2017 at Betabit Utrecht

Ähnlich wie Global Azure Bootcamp 2017 at Betabit Utrecht (20)

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Global Azure Bootcamp 2017 at Betabit Utrecht

Hinweis der Redaktion