Java Map Notes , Immutable objects and its importance,Strategy pattern

 HashMap allows null as a key and so hashcode for this will always be zero , even empty string can be given as a key and it also hols the hash code zero but internally they  are stored in different indexes

Concurrent hashmap does not allow null as a key , Hash table also does not allow null key

Some sample example below

Oct 06, 2020 12:39:06 PM MapContainsGet logMapInfo

INFO: HashMap: {NEVADA=Carson City, WASHINGTON=Olympia, NORTH_DAKOTA=Bismark, MONTANA=null, ARIZONA=Phoenix, UTAH=Salt Lake City, OREGON=Salem, COLORADO=Denver, SOUTH_DAKOTA=Pierre, IDAHO=Boise, NEW_MEXICO=Sante Fe, CALIFORNIA=Sacramento, WYOMING=Cheyenne}

Oct 06, 2020 12:39:06 PM MapContainsGet demonstrateGetAndContains

INFO: 

Map of type java.util.HashMap returns null for Map.get() using Montana

Map of type java.util.HashMap returns true for Map.containsKey() using Montana

Map of type java.util.HashMap returns null for Map.get() using Kansas

Map of type java.util.HashMap returns false for Map.containsKey() using Kansas

Oct 06, 2020 12:39:06 PM MapContainsGet logMapInfo

INFO: LinkedHashMap: {ARIZONA=Phoenix, CALIFORNIA=Sacramento, COLORADO=Denver, IDAHO=Boise, NEVADA=Carson City, NEW_MEXICO=Sante Fe, NORTH_DAKOTA=Bismark, OREGON=Salem, SOUTH_DAKOTA=Pierre, UTAH=Salt Lake City, WASHINGTON=Olympia, WYOMING=Cheyenne, MONTANA=null}

Oct 06, 2020 12:39:06 PM MapContainsGet demonstrateGetAndContains

INFO: 

Map of type java.util.LinkedHashMap returns null for Map.get() using Montana

Map of type java.util.LinkedHashMap returns true for Map.containsKey() using Montana

Map of type java.util.LinkedHashMap returns null for Map.get() using Kansas

Map of type java.util.LinkedHashMap returns false for Map.containsKey() using Kansas

Oct 06, 2020 12:39:06 PM MapContainsGet generateStatesMap

SEVERE: java.util.concurrent.ConcurrentHashMap does not allow for null values - java.lang.NullPointerException

Oct 06, 2020 12:39:06 PM MapContainsGet logMapInfo

INFO: ConcurrentHashMap: {NEVADA=Carson City, WASHINGTON=Olympia, NORTH_DAKOTA=Bismark, ARIZONA=Phoenix, UTAH=Salt Lake City, OREGON=Salem, COLORADO=Denver, SOUTH_DAKOTA=Pierre, IDAHO=Boise, NEW_MEXICO=Sante Fe, CALIFORNIA=Sacramento, WYOMING=Cheyenne}

Oct 06, 2020 12:39:06 PM MapContainsGet demonstrateGetAndContains

INFO: 

Map of type java.util.concurrent.ConcurrentHashMap returns null for Map.get() using Montana

Map of type java.util.concurrent.ConcurrentHashMap returns false for Map.containsKey() using Montana

Map of type java.util.concurrent.ConcurrentHashMap returns null for Map.get() using Kansas

Map of type java.util.concurrent.ConcurrentHashMap returns false for Map.containsKey() using Kansas

Oct 06, 2020 12:39:06 PM MapContainsGet logMapInfo

INFO: EnumMap: {ARIZONA=Phoenix, CALIFORNIA=Sacramento, COLORADO=Denver, IDAHO=Boise, MONTANA=null, NEVADA=Carson City, NEW_MEXICO=Sante Fe, NORTH_DAKOTA=Bismark, OREGON=Salem, SOUTH_DAKOTA=Pierre, UTAH=Salt Lake City, WASHINGTON=Olympia, WYOMING=Cheyenne}

Oct 06, 2020 12:39:06 PM MapContainsGet demonstrateGetAndContains

INFO: 

Map of type java.util.EnumMap returns null for Map.get() using Montana

Map of type java.util.EnumMap returns true for Map.containsKey() using Montana

Map of type java.util.EnumMap returns null for Map.get() using Kansas

Map of type java.util.EnumMap returns false for Map.containsKey() using Kansas

Best link to read

https://www.devdiaries.net/blog/Java-Recruitment-Questions-HashCode/

Immutable objects are a very important element of application building, which makes applications more legible and consistent, as well as more error-proof. Every developer should know how to use immutable objects and what benefits are associated with their use.

Writing this post I didn’t think about the disadvantages of creating immutable classes. Fortunately, a user potracheno informed about it in the comment. Thanks!

Immutable objects have far more advantages than disadvantages. The only thing that comes to my mind is the cost of memory. With immutability, any time you need to modify data, you need to create a new object. This can be expensive. If you are sure that you will not need immutability (e.g. you are creating a simple single-threaded application) - do not create immutable code.

Advantages of immutable objects

These objects make us avoid accidental changes, often in very inappropriate places. If an object is changeable, there will definitely be someone who wants to change it where it should not.

A good example of such a situation is the object that we pass as a parameter of the method. Such an object can be passed between multiple application layers, between multiple method calls. It can be passed on very deep in the call hierarchy. This makes it very difficult to identify where it was changed. This can lead to many strange and difficult to solve problems.

Using immutable objects we do not have such problems and the design of our application improves.

Invariant objects are also safe for multithreaded use. If an object is not changed, it can be safely transferred between threads without synchronization.

Another advantage is that such objects are ideal as key objects in maps. In a situation where keys are variable, after changing the key object its hashcode changes, making it impossible to find the stored value in HashMap.

How to create immutable objects?

Such objects can be created in three ways. Through the constructor and that is the easiest way. But it has a fundamental disadvantage. The more fields to initiate, the more parameters in the constructor. That’s why you shouldn’t create objects that have more than 2-3 fields.

          
public final class Animal {

    private final String name;
    private final String ownerName;

    public Animal(String name, String ownerName) {
        this.name = name;
        this.ownerName = ownerName;
    }

    public String getName() {
        return name;
    }

    public String getOwnerName() {
        return ownerName;
    }
}

// Creation
Animal animal = new Animal("Tina", "John");

Another way is the factory method. As with the constructor, the more fields, the more parameters. But this approach has such an advantage that we can create several such methods with a different names, with different set of parameters, which improves readability.

A third way to create immutable objects is to use the builder pattern. In order to use the builder, it must be implemented inside the class so that it has access to private class fields.

          
public final class User {
    private int age;
    private String firstName;
    private String lastName;

    public static Builder builder() {
        return new Builder();
    }

    public static class Builder {
        private User user = new User();

        public Builder firstName(String firstName) {
            user.firstName = firstName;
            return this;
        }

        public Builder lastName(String lastName) {
            user.lastName = lastName;
            return this;
        }

        public Builder age(int age) {
            user.age = age;
            return this;
        }

        public User build(){
            return user;
        }
    }

// Creation
 User user = User.builder()
                .firstName("John")
                .lastName("People")
                .age(45)
                .build();

We can write such a builder manually or use some kind of IDE plugin to generate it for us. Another option is to use Lombok and @Builder annotation.

What is an immutable object?

An immutable object is one which, after being created and initialized, remains unchanged and, importantly, cannot be changed (in a conventional way, of course). This means that such an object does not provide methods that allow changing its state. And all its fields are private (or public final).

This changeability is a little apparent, because, using reflection, you can change the field values in each class. However, this does not mean that for this reason, we should give up immutable objects.

How to ensure the object’s immutability?

Apart from the fact that all fields should be private, they should be final. And the class itself should be marked as final so that it cannot be inherited. Such an object should not provide methods to modify its internal state, e.g. setters.

But private, final fields are not everything. It is important what types of data we store in these fields. It should also be unchangeable! If the fields in your object are primitives, primitive wrappers or Strings then there is no problem. All these types are immutable. But if you use your own objects, you must ensure that they are unchangeable.

https://www.devdiaries.net/blog/Java-Recruitment-Questions-Immutable-Objects/

Strategy defines a family of interchangeable algorithms (in the form of classes) that are used to solve the same problem in several different ways.

Number converter with the strategy pattern

Imagine that you want to create a system to convert numbers into different Numeral Systems .
As in the previous example, we will start with the interface:

      
public interface ConvertingStrategy {  
    String convert(int number);  
}  

Time for implementation. We will add support to convert number to octal, binary and hex system:

      
public class BinaryConverter implements ConvertingStrategy {  
    public String convert(int number) {  
        return Integer.toBinaryString(number);  
    }  
}  
  
public class HexaConverter implements ConvertingStrategy {  
    public String convert(int number) {  
        return Integer.toHexString(number);  
    }  
}  
  
public class OctaConverter implements ConvertingStrategy {  
    public String convert(int number) {  
        return Integer.toOctalString(number);  
    }  
}  

According to the diagram at the beginning of the post - we will create a class Context that will use our strategies:

      
public class Context {
    private ConvertingStrategy strategy;

    public Context(ConvertingStrategy strategy) {
        this.strategy = strategy;
    }

    public String convert(int number) {
        return strategy.convert(number);
    }
}

An example of how to use a strategy to convert:

      
public class Main {  
    public static void main(String[] args) {  
        Context ctx = new Context(new HexaConverter());  
        System.out.println(ctx.convert(1000));  
//      Result: 3e8  
//      If you change HexaConverter to BinaryConverter, the result will be: 1111101000  
    }  
}  

https://www.devdiaries.net/blog/Java-Strategy-Design-Pattern-By-Examples/

https://www.devdiaries.net/blog/Java-Interview-Questions-Multithreading/

https://www.devdiaries.net/blog/10-Object-oriented-design-principles-everythone-should-know/

Best Practises while using Collections

----------------------------------------

Always use interface type as a reference type.
Always use interface type as a return type
Always use Interface Types as a method argument
Use generic type and diamond operator
You can declare a collection of a generic type like this:
List<Employee> listEmployees = new ArrayList<Employee>();
Since Java 7, the compiler can infer the generic type on the right side from the generic type declared on the left side, so you can write:
List<Employee> listEmployees = new ArrayList<>();
 Prefer isEmpty() over a size()

Return empty collections or arrays, not nulls
Do not use the classic for loop.Use ForEach Loop which uses iterator behind the screen
Overriding equals() and hashCode() properly

Using Arrays and Collections utility classes

Prefer concurrent collections over synchronized wrappers
When you have to use collections in multi-threaded applications, consider using concurrent collections in the java.util.concurrent package instead of using the synchronized collections generated by the Collections.synchronizedXXX() methods.

It’s because the concurrent collections are designed to provide maximum performance in concurrent applications, by implementing different synchronization mechanisms like copy-on-write, compare-and-swap, and special locks. The following list shows you how to choose some concurrent collections (on the right) which are equivalent to the normal ones (on the left):
- HashMap -> ConcurrentHashMap

- ArrayList -> CopyOnWriteArrayList

- TreeMap -> ConcurrentSkipListMap

- PriorityQueue -> PriorityBlockingQueue

Certification Learnings

AZ - 204T00A-A -->Microsoft Certification name

Certifications focus areas

--------------------------------------

Azure funtion is server less architiecture

Logic Apps -->

 https://www.skillpipe.com/  azure@123 and uname is my off mail id

pp Service

------------------------------------

PAAS Service -->

Users location should be region in resource group

App service plan question

-------------------------------

this is nothing but a kind of subscription plan which is used to host a website

ACU -- Azure Compute Units(cmp 2 plans with this)

A-series(basic virtual machines with in cloud) compute equivalant

Is there a Limit number of webapp (created) can run under app service plan?

No restrictions put by azure

Basic plan incurs cost anly dev and test has free

Automatic scaling not possible in basic plan

Scaleup/Vertical Scaling -->can be Done manually only(change priciing tier) (from B1 tp Premium example)

Scaling out / horizontal scaling --->

A single PaasVM is not sufficient i.e getting a lot of users , usage performance gone up then azure gives auto scaling option in this case .Azure creates replica of the same VM  what web site is deployed in first paasVM will be deployed in the new VM also.

Autoscaling used when load increasing.Load balancing will be taken care by azure

Horizantal scaling and auto scale both are sames

when load reduces again due to business reasons in the future post auto scaling, how azure handles those cases?

We can define rules(like when traffic is more than 80% then start auto scaling and when less than 60% do not auto scale)

Encrypted credentials in publishing profiles

We can download profile from azure portal(profiles like dev, UAT,Production)then using this profile we can publish from IDE

for Azure Service plan by default it won't create a storage account?

It won't create.

Azure web app can connect with onPremise data base

Yes, this is called hybrid approach

Options 

1)Networking tab Azure provides

Virtual Network configure(Data base will be in this n/w)

2)Hybrid connections

with help of gateway/connection manager

One - One configuration

https://docs.microsoft.com/en-us/azure/app-service/app-service-hybrid-connections

for each app service plan we have a dedicated PAAS VM ?

Depends on the configuration.If basic then 1 PASSVM

any limitation how many web apps can be deployed within a resource group?

No such limitation

from lokesh reddy to everyone:

Logging and tracing enabled through App Insights

we have classess for logging

Azure Traffic Manager(Certification )https://docs.microsoft.com/en-us/azure/traffic-manager/traffic-manager-faqs

-----------------------------------------

It's a load Balancer DNs based.Its on a global level we want to do.

What is DNs based ??

Refer DNS based.jpg

DNS Level Load balancing is done by Azure Traffic Manager

We can enable traffic manager in Azure portal

This is recorded while doing task imgapinamratha.azurewebsites.net

Deployment slot

-----------------------------

Down time reduces(SWAp from dev to UAT)

What is happening under the hood in swapping

--------------------------------------------

Virtual IP swapping is the thing Azure is doing(Changing the IP reference)

16/6/2020

------------------------------------------------------

Storage Accounts

---------------

Blob storage 

unstructured data(video,image file) stored in this blob storage

Is Blob storage in some way related to HDFS (Hadoop distributed file system) storage?

500 Terra byte -- Single storage account

Classic (w.r.t) Azure portal

Modern azure Portal  -- ARM -- Azure Resouce Manager API

Classic Portal  -- ASM -- Assure Service Management API

LRS --->Replication --Durability of the data

Access tier - hot/cool Active data /Passive data which we are storing

Blob storage 

Block blobs  -- Videos,PDF (100 mb chunks)  (Single file 4.75 tb can be stored as a single block blob in block blobs,A single block blobs contain 50000 blocks)

Append blobs  --- Composed of blobs.Optimised for append operations.Each blob can be different type 1.95 Gb maximum size of single Append blob.Editable file.Text logs.Append operation

Page blobs --- storing virtual hard disk.Similar to harddisk storage

Az - copy is tool to upload and download huge amount of data

Durability options

--------------------

LRS (Lease Redundant storage)-- cheapest option.Create 3 replicas of data.Protect against rack failures,disk node

Refer snapshot durability.This data center fails then gone

ZRS (Zone Redundant storage)--Data replicated synchronously.3 replicas , 3 zones,one region.Entire region goes down .Protect against rack failures,disk node,zone failures

GRS  Multiple Regions  -- six replicas  , 2 regions(3 per region).Async copy to secondary.protect against major regional disasters.

RA - GRS

-------GRS + Read access to Seconday.We can use both the copies.In normal GRs till primary available we cannot use secondary

-------GRS + Read access to Seconday.We can use both the copies.In normal GRs till primary available we cannot use secondary

GZRS -- Refer Durability_1 snapshot

RA-GZRS -- Refer Durability_1 snapshot

Blob level we get archive option

Shared Access Signature key

https://docs.microsoft.com/en-us/rest/api/storageservices/create-account-sas?redirectedfrom=MSDN

SAS token even private then we can access 

SAS tokens can be grouped together

Blob storage Client libraray helps us to create storage Account and maintain Our data

Exclusive access for modifying a blob

Then take a lease (lease.png snapshot)

Azure queue storage message can remain for 7 days

CosMos DB

------------------------------

What is no SQL Data base

If we need to display data like face book then using relational database we need to join many tables...Not a feasible approach

Problem with relational data base is predefined scehma like employee record in employee table and so on

RDBMS limits

--------------------------

It cannot go on scaling i.e no horizonal , vertical scaling

Sharding pattern if we implement then it can horizally scale the data base

In no SQL we do not have any schema

Then how we store data?

Why no SQL

Store infinite data with out performance issue,wanted hyper scale

Brain storm on how we create partition

Plenty of No SQL databases

-----------------------------

MongoDb,Casendra then why cosmos? COSMOS is a kind of wrapper

Internally uses MONGO DB

Using COsMOs we can do global replication

Consistency is a challenge?How to overcome this?

Social media platforms can use loose  consistency.Advantange is that application/database is highly available

On the othet hand there are applications which deal with transactions then consistency is imp then availability will go down

Certification will have consistency level questions

https://docs.microsoft.com/en-us/azure/cosmos-db/request-units

17/6/2020

------------------------------------------------------------

Server less services/Architecture

What is azure Function

-------------------------------

Serverless component

web App and a Function App difference

------------------------------------------------

Web App -->If we change a function in the code then we need to deploy entire web App again.It should be one language

With Micro services coming in picture its a new architecture , in which we try to break application to the most micro level

Example Micro services amazon web site send an email to user number of items purchased(1 micro service)

What is the advantage of micro service is 

Testing of entire is not needed only changed functions can be tested

Each of the function can be written in different language

Each of the Azure function can be hosted as Micro service

To start an Azure function

Call it from http end point

Logic App can call function App

Azure Kubernetes service helps to write Micro service

Durable functions

---------------------------------

Gives a specific design pattern

Used in case of chaining  F1-->F2-->F3

Refer Durable snip shot

https://github.com/Azure/azure-functions-durable-extension/

It is an extension of azure function

Orchestrator function -> Checks the health of all the functions which are in chaining

Durable Function Scenarion  Fan In/Fan Out -->Multiple functions in parellel

What is a Logic App

-------------------

Its a kind of business logic layer.Something like middle layer.Its a designer focused tool which azure is giving to do integration(like intergration of services)

Logic_App SnapShot

We can increase number of instances

We can create work flows inside logic app

Custom connectors can be written in logic apps

https://docs.microsoft.com/en-us/connectors/custom-connectors/create-logic-apps-connector

IAAP

----------------

ARM Template

In Azure we can create different virtual machines

Check what is hyper-v? will have .vhd extension

MAnaged / Unmanaged disks

-------------------------------------

Each of storage account has limitation of In putOut put operations

Looking for higher IOP's then managed disk is risky

Microsoft spreads across various storage accounts so that we get higher IOP's

Highly available env azure provides??

Availability set if you put 2 machines then it will meet 99% assured SLA

fault domain,update domain

Dockers , Containers

============================

Dockers are for quick deplyment

Difference b/w container and virtual machine

Each of Virtual machine has its own OS

Container is a isolated componenet for running application 

Container can run on any operating system

Dockers help you to create containers

https://docs.docker.com/engine/reference/builder/

Azure Container Registry

------------------------------------------

To maintain code base

This is build on Docker registry service

ACR Build

-------------------

Helps to stream line whole process of Build process container

Kubernetes is the famous orchestrator developed by google

-----------------------------------------------------------

orchestrator is some one who will run the show

-------------------------------------------------------

Since containers are there we need orchestrator

AKS (azure Kubernetes service)(Cross vendor maintanace is the advantage)

If we do not want to use AKS then use Azure Container Instance(Fastest and simplest way to run a container in Azure)

18/6/2020

-----------------------------------------------------------

Implement User Authentication and User Authorization

--------------------------------------------------------------

Authentication --- CGI Id card to enter office premise

Authorization -- What we can do once we are in?

Microsoft Identity Platform which provides Microsoft authentication library

Microsoft graph : common api to work with all 0365 resources as well microsoft with all microsoft suit(like access calendars,..etc)

Azure Active Directory : Runs entire show in terms of Authen and autho

Local AD?why we cannot use?

SSO -- Single sign ON what is?

Sync pws when we sync local AD  to Azure AD?

It is our choice of sync pwds

ADFS we ue when pwd sync

-----------------------

Active Directory Federation service

Authenticate user against remote directory

SAML tokens generated as a part of ADFS Authentication

Active Directory Authentication Library(ADAL)

------------------------------------------

Microsoft Graph

--------------------------------

Rest APi Endpoint that helps to work with different services

Graph explorer tool gives the end point

https://developer.microsoft.com/en-us/graph/graph-explorer

MSAL -- Microsoft authentication library

https://github.com/MicrosoftLearning/AZ-204-DevelopingSolutionsforMicrosoftAzure/blob/master/Instructions/Labs/AZ-204_06_lab_ak.md

All data (including metadata) written to Azure Storage is automatically encrypted using Storage Service Encryption (SSE).

Implement Secure Cloud Solutions

------------------------------------------

How to secure credentials --Azure KeyVault(Its a dedicated pass service where it can manage certificates,credentials)

HSM -- H/w security modules these are backed by Azure Key Vault

Each of the vault will have a owner assigned

Authentication

Managed Identity(https://docs.microsoft.com/en-us/azure/active-directory/managed-identities-azure-resources/overview)

Service Principal

Entity which will have some sort of permission associated with it

it can be pwd based

it can be certificate based

API Management

-------------------------

PAAS Service that microsoft is giving for management of API

We can use this with existing APIs no need to make any code changess

Working components

Certification question regarding policy adding in API management

19/6/2020

------------------------------------

Message

Azure event grid ,event hub,notification hub

Event grid:

service which listens to events

WebHook : Applications are using this as an event's when ever this happens we have incoming and outgoing webhooks

When ever some specific item in cosmos db is updated then do some action this can be achieved using event grid

Sunbscription can be created with filters and based on this we can filter events

Azure Event Hubs

-----------------------

Specialized for huge amount of data

It is a streaming service

It can take millions of events per second

Take a burst of dta from different sources then plug event grid and then pass it to subscribers

With in Event hub we have paritions

Notification Hubs

---------------------------------

App Developers

PNS -- Platform notification system 

Messaging options

--------------------------------

Specialised for highly reliable messaging

----------------------------------------------

Service BuS and Queue storage:

These are Queing service

1 MB size of message in a queue

FIFO

3 communication mechanisms

Queues

Topics

Relays

Message time to live is not asked in event grid but asked in queue,topics

Queue Storage

-------------------------------

It is with the storage account

Messages can be out of sequence

Single msg can be delivered multiple times

event grid domain helps you manage a large amount of event grids that are related to each other https://docs.microsoft.com/en-us/azure/event-grid/event-domains

Instrument Soulutions to support monitoring and logging

--------------------------------------------------------------------

Monitoring services

Azure Monitor -->Connect data from variety of servers

We can have alerts created

Trouble shooting steps can be done through azure monitor

Log Analytics

Metrics

We can create Alert rule in monitor

which lang is used for querying here in Logs

Kusto

Application Insights

--------------------------------------

Feature of Aure Monitor(APM Application performance Management service)

Instumentation key we can put and check logging in Azure portal

Load testing can be done from Application Insights

Its  a stand alone servce can be used for non-Cloud service also

Implement code that handles transient faults

-----------------------------------------------------------

a temporary issue

Code that we are writing in a cloud shold be good enough to find any transient issue

Retry

CDN

Content delivery

Amazon, netflix has CDN centered architecture

Purging

---------------------------

Flush down across the servers

Azure cache for redis

------------------------------------

Distributed cache technology is provided by redis

Distributed cache?(inmemory datasturcure)

Multiple instances of Web Apps

Web front end server and each of which will have its own cache

refer distributed cache.png

VM1

Vm2

VM3

My Learnings from browsing

-------------------------------------------------------

 Only storage accounts of kind StorageV2 (general purpose v2) and BlobStorage support event integration. Storage (general purpose v1) does not support integration with Event Grid.

Reference:

https://docs.microsoft.com/en-us/azure/storage/blobs/storage-blob-event-overview

Learnings

-------------------------

Linux os is the only one that supports container instances to be available from virtual n/w

Millions of aggreements per hour then use data ingestion service like Azure Event hub.You can use event capture to store aggrements in to azure blob storage for long term storage

We can use lease ,so that no other process can modigy the blob while the blob is being accessed

On-Premise data gateway to allow logic app to access onpremise data resources

For MongoDB ,if we need migration with minimum downtime , we need to implement online migration using Azure database migration service

Accelearated networking to improve performance of Azure VM

Authorisation level - Function -->here authorisation key is sent  with the function invocation .This adds extra level security for azure function

Deplying webapp to VM is Infrastructure As A service(maintananace is required)

Deploying web App to App service Plan is Platform as service(internally web app deplys to VM)--->no maintanace is req--underlying infrastucture management not required

Configure back up to create restorable archive copies of your apps content ,configuration,database.Buthere App service Plan should be standard or higher

App service plan can be scaled up at any point of time

Std app service plan and premium app service plan required for deployment slot option

Azure web job can be used to run a background task and this is a part of web app

Azure function should have a storage account

Azure web job can be used to run a background task and this is a part of web app

.NetCore 3.1 is supported on both windows and linux based platforms

ASP.NETV4 is supported on windows platform only

Blob storage also can store VM disks

If we use AZ-copy tool to upload a blob to container then here we have to give permissions to user using Access Control in Storage Account

We can add meta data to the blob in the container and we can use this to fetch the container data

Using Shared access Signature we can access blob with pvt access.Here link is generated to access blob and has a specific time interval with in which we can access,we can give specific ips only ,can specify expiry data time.

Sahred access signature can be at blob level , or at account level.using Share access signature we can give read , write i.e specific permissions can be given.

Sanp Shot at blob is nothing but creates a copy of blob item

Lease operation to acquire lock on blob so that no other client can access it 

Hot spot when data is inserted in to only one partition 

Function App host.json -->we can specify time out here

   -->Consumption Plan valid range is 1 sec to 10 minimum

   ---->Premium plan it is 1 sec to 60 minimum

   ----->Dedicated App service Plan there is no overall limit

we have retry support by default for Azure Blob Storage,Azure Queue Storage , Azure Service Bus(queue/topic)

For Azure queue storage and service bus storage if retry fails messages are sent to poison queue

In case of consumption plan for Function App , always ON setting  option is not there.It is available for App Service Plan

YAML file/ARM template can be used to deploy group of containers

Once partition key is set to the container , it cannot be changed

Change feed --> listens to all changes in the container and gives the list of changes

Azure Cosmos DB

Storage procedure

Trigger(Wehn adding data to container we can append some data)Pre trigger

Function trigger(chnage feed)

Synthetic Partition Key

---------------------------------------

Does not have any propert ideal for partition key

As a work around we can do the following -->combination of some fields/ one field+random suffix

---------------------------------------------------------------

Time to live can be set on the items in the container or at the container level

Order by the items like asc then should be a part of composite policy, they should be defined in indexing policy

Access control can be added can resource group level or at indivisual resources or at subscription level also

Owner role

Manage every thing including access to resources

Contributor role

Manage every thing except access to resources

Reader role

Response type(Mentioned by client) as code then it is authorisation code work flow with external service provider which the user is using

What happens in autorisation work flow?

---------------------------------

1)Client ----> External Identity Provider (Goolge)(confidential clients)

2)Goolgle--->(responds with authorisation Code)- to Client

3)Clinet-->(again sends this authorisation code to the google so that it gets access token) google

4)Google gives access token so that now client uses this and access the resorce server from which it needs data

Above steps of exchanging authorisation code to get access token are done using back end channel

Java script applications can use implicit workflow where there is no authorisation code directly it asks for access tokens

external identity provider will have two uls one to give authorisation code and other to give access tokens

OATH 2.0 implicit workflow(Here the response type is token)

----------------------------------------

Single page application(Called as public clients)

Not hosted on server

Client Credentials work flow

---------------------------------------

Used by the clients to obtain access tokens outside context of the user i.e we do not have user here

OPENID Connect

------------------------

used only for autentication on top of OATH.response type is code.when we exchange authorisation code we get id token

Microsoft Authentication Library used by developers to get tokens

Multifactor Authentication in Azure AD

----------------------------------------

Conditional Access Policy -- we need Azure AD Premium p2 license is needed

Group claims

-------------------------

we need to modify application manifest file

Azure Key Valult

------------------------------

To store secrets

Service Principal

-------------------------------

Used by external application to access the resources(secrets,certificates,keys) in azure key vault

Create this service principalin azure ad using power shell commands from there you get application id, tenant,password

set the env variables

execute one more command to set policy so that service principal gets permission over secrets

Role based access control given at azure key vault level

Data in storage account can be encrpted by using a managed key , this is done by azure it self

Managed service identity -- can be assigned to storage account so that it can authorise it self with azure key vault service and get the key and use it for encrption

Auto Scaling

----------------------------

Conditions can be written based on auto scaling increasing VM instances can be done

Scale up is used incraese tier of App service Plan

Scale out is used to enable auto scaling i.e increase VM

Application Insights

---------------------------

funnels (user going from index page to products page)--> All pages whether user is going through

UserFlows

Impact

Retention(users are returning back to your application)

Availability feature is also part of application sights

Ping test we can do as a part of this.From different locations we can do ping test

APPlication Telemetary to add application insights

Custom telemetry is also possible.

Azure cache for reddis

--------------------------------------

Fetching of data is much faster as all of data is stored on RAM

Massive amounts of transational data should be stored on sql / no sql data store should not store on azure reddis

Frequently accessed data should be stored in Azure Reddis

Different types of data can be stored in azure cache for reddis

Data which is changed every min then it should not be stored in reddis

Application should ensure that data is updated in  reddis cache

In reddis data is stored in form of key value pairs

Tiers available here also

Premium tier --> data stored in clusters,data can be stored on some disk

SetString to set data on the cache

Stringset method to set key/value pair

Data Invalidation in cache

--------------------------------------

Some times data in SQL D is updated and so cache should alos be updated so in this case update cache or invalidate your cache

Invalidate -- remove key all together and this is manual effort

We can add expiry time/policy to the items in cache  ( we have a class DistributedCacheEntryOptions which has expiry time method)

stack exchange for redis package for console based application here we have IDataBase

Content delivery N/w

------------------------------------

To reduce latency

Here we have point of presence ,and edge servers

Edge servers are located globally

Uers will make a call tp Azure content deliver n/w service then it will see  user location and then transfer request to that edge server near to user.

Edge servers can also cache data.Edge server will check with POP if POP has response then sends to the user else redirects request to origin and then the response is cached by POP for future use

Using this users across the world will have same experience with the web app

This also has different pricing tiers

Caching in Azure CDN

------------------------------

BypASS  --- Do not cache even application is setting some cache

OVERRIDE

SETIF MISSING

Cache in case of query string

-------------------------------------------

ByPass Cache

Cache every unique url

Ignore query strings

Azure Front Door Service

-------------------------------------

client req are sent to most fast performing backend

Back end pool can be web app

Weightage,Priority can be specified to each back end

health probes we have here which keeps checking our backend service

File size should be b/w 1 KB and 8 MB, supports all compressions brotil and gzip

Transient faults

------------------------------

Azure is a shared env

We can do retry in these cases

Azure service bus

---------------------------

Messaging service

Queue service

FIFO message delivery

Topic service

Multiple consumers(subscribers)

When we create instance if service bus we are creating name spaces

Here in basic tier we do not have topic from standard tier we do have(Tier can be upscaled at any point in time)

We create name space then either queue/topic can be created

using service bus explorer we can send a message to queue

we have two operations in queue

peek  (Active message count remains same)

receive(Active message count decreases)

Queues/Topics have their own shared Acess Policies where we give permission of sending message,receiving msg,listing mesg

Message sent on queue should be in bytes.SendAsync Method we use

Receiving msg two step operation

receive and process message

delete msg 

Complete async will do the above two steps

We have message handler which gets triggered when we receive message from the queue

Difference b/w storage queue and this queue service is in storage queue we have to get count of messages which we have received and then we display it but in this queue Message handler is triggered when message is reveived

Properties

Time to live(deleted from main queue once time expires)

Dead letter queue contains msg that cannot be processed/ did not deliver to any receiver

Lock duration --> default is 30 sec(gets invisible for 30 sec)

peek message ability to see

Receive and delete

Message object

broker properties -- Content type we are sending in msg , message id(unique number given from application perspective this can be used in case of duplication detection feature),sequence number(64 bit int assigned on msg),correlation id(replier scenario),session id, Reply to session id 

Duplicate detection

-------------------------

can be enabled in azure in queue section, msg id can be given unique

Corelation id advantages

-----------------------------------

Correlates multiple messages together

Event Grid

----------------------------

Event sources (publishers) Azure blob storage , Custom Events, Resouce Groups

Event handlers -- event hubs, functions ,logic apps etc

event Grid Schema

---------------------------------

Event emitted from a aprticular resource follow a particular schema

events are sent from source to event grid as an array and size of array is 1 MB

Azure service bus can also be a receiver to the events emitted from an event grid

event grid also has filters

Particular event has a subject property we can add filter by this like subject ends with and subject begins with

we have custom filters also like if data length less than 12 bytes then receive events

Features for subscriptions for events

------------------

max event delivery attempts 30 -->try 30 times to deliver event to end point

dead lettering

event subscription time

Custom web hook as event handler

-------------------------------------------

Handshake is needed with event grid at first

In Hand shake event grid sends validation url and validation code to our custom web hook

Custom web hook also responds with the same

nrock tool to expose web applications running in local

Azure functions as web hook te receive events from event grid

--------------------------------------------------------------------------

Custom topics

-------------------------

here event which we send from custom resource to custom topic should follow event grid schema

Azure Event Hubs

-------------------------

Big data streaming platform and event ingestion service .It  can receive millions of events per second

Receive events from multiple devices then we can use this server less service (no need to manage infrastructure)

Not a persistent data store

It can route the data to Azure blob storage

First create a name space

then we can create multple event hubs

Multiple partitions here we have so that through put is great

Once the message is read it is not deleted , Message retention should be specified

Minimu is 2 partitions max is 32

Msg can be sent on one partition also

readers can also read from particular partition

Each message will have a unique sequence number

Consumer groups

------------------------

we have multiple consumers groups for event hubs

Recomendation is one active receiver on a partition per consumer group

Offset

------------------------

Used so that same messages are not read again and again

AzureMessagingEventHubProcessor keeps track of the messages that have been read from the event hub

Update check point method helps in tracking the above point

Azure search service(AZ-203)

-------------------------------------

Here we use index

Index is like a table and documents are like rows in the table

Only retrivable attribute can be changed at any point of time.Once the index is in place no other attribute other than retrivable can be changed

Azure API Management service

--------------------------

iNTERFACE B/W USERS AND BACK EnD api

API Policies we have -- xml operations

CheckHttpHeader

Limit call rate

Authentication policy

We have conditions with in policy

Application can afford read out of order writes so consistency can be eventual

Cache 

security for 

OPEN API SPECIFICATION

---------------------------------

swashbuckle checls for the specification

SWAGGWER API