README

	Overview
	========
	
	Backup service for TodoServer is a java application with REST web services. 
	System allows backup data from TodoServer. Backup process is fully async, 
	any backup requested is done produced in background and can be retrieved.
	
	
	
	Dependencies
	============
	To be able use backup service you will need:
	- Java installed, JDK at least 1.8
	- Maven installed
	
	
	
	Package content
	===============
	Server package consists of:
	- Project source folder: src
	- Maven build file: pom.xml
	- Windows script files: run.bat, test.bat
	- Linux/Mac script files: run.sh, test.sh



	Build & Run
	===========
	Backup server build is fully powered by Maven. Build sources, run application, 
	run integration tests you can do by running maven with different parameters:
	- Cleanup: 			mvn clean
	- Compile sources:	mvn compile
	- Build jar: 		mvn package
	- Run application: 	mvn spring-boot:run -DskipTests=true
		^^ requires compiled sources
	- Run tests:		mvn test
	
	Additional scripts are provided to simplify usage:
	- Windows
		- run.bat - build sources and run applicaton
		- test.bat - build sources and run integration tests
	- Linux/Mac
		- run.sh - build sources and run applicaton
		- test.sh - build sources and run integration tests



	Service configuration
	=====================
	Service configuration is stored in the property file:
		src\main\resources\application.properties
	
	Detail information on configuration is available within this file.



	Project information
	===================
	- Backup server is written on java
	- Maven used as build tool
	- Backup server backups data from TodoServer, information about TodoServer is configurable
	- Backup server provides REST endpoints
	- Backup server will not use persistent storages (at least this version)
	- Backup server works asyncrhonous from client perspective	
	- Backup server can work even if TodoServer is not running, however backups will not be produced

	
	
	Technologies
	============
	- Spring - dependency injection system
	- Spring Boot - application container with powerfull set of features
	- Swagger - Displays information on REST endpoints	
	- Jackson - Json processing utilities
	- JUnit - Code testing framework
	
	
	
	Solution
	========
	Client using REST endpoints is able to schedule backup for execution.
	All backups may be executed one by one or simultaneously depending on 
	configuration. When backup is scheduled its id is added to 'backup queue'
	for processing and new backup status record is created. Backup queue 
	dispatched in separate thread, on this stage server initialize all
	required storages and retrieves list of users from TodoServer. 
	Server schedules backup processing for each pair: backup id and user id.
	Backup of each pair is done in async way.
	During backup of the account server retrieves all user Todo's from
	TodoServer, all of them are stored into local storage.
	If backup is run into issues all non-processed tasks for the same backup will be
	skipped no more new ones will be created. 

	
	
	Development assumptions
	=======================
	- To improve usability backup status will include 3 different dates:
		- date requested - date when backup was scheduled
		- date started - date when backup was started 
		- date finished - date when backup was finished (successfuly or not)
	- Server limit amount of queued backups to avoid abusing
	- Server limit amount of concurrently running backups to avoid memory issues
	- Server limit amount of concurrent http requests to TodoServer to prevent abuse
	- Server will return error if client will do export request in these cases:
		- backup does not exist
		- backup is in progress
		- backup failed
	- Server will mark backup failed when:
		- It can't retrieve list of users
		- It can't retrieve data for any user
		- It can't store any todo data
	- Server will not backup users without todo items

	- Server will not use async messaging - with such amount of features it will add extra
		code complexity, but would be nice to have when more features will be added
	- Server will not share any data between different backups - this may significantly 
		improve backup performance but limit ways to extend server features. Also
		It kills the idea of multiple backups: all backups will have the same data,
		retrieved at the same point of time.
	- Server will not try to finish already failed backups - client can check status
		for each backup and schedule new one in case when previous was failed. If server
		retry backup and user will request new one we will have two similar backups



	Public resources
	=======================
	- Backup accounts (rest endpoint)
		- request: POST /backups
		- request body: N/A
		- response body (json):	
			{
				"backupId": <backup Id>
			}

	- List backups (rest endpoint)
		- request: GET /backups
		- request body: N/A
		- response body (json):	
			[
				{
					"backupId": <backup Id>,
					"dateRequested": <date backup was requested>,
					"dateStarted": <date backup was started>,
					"dateFinished": <date backup was finished>,
					"status": <backup status>
				},
				...
			]
			
		Supported backup statuses:
			- In progress
			- OK
			- Failed

	- Export backup (rest endpoint)
		- request: GET /exports/{backup id}
		- request body:	N/A
		- response body (csv):	
			Username;TodoItemId;Subject;DueDate;Done
			{userName};{todoItemId};{todoItemSubject};{todoItemDueDate};{todoItemDone}

	- Display endpoints information (html)
		- request: GET /swagger-ui.html
		- request body: N/A
		- reponse body : <html>
	
	Note: with default configuration these resources will be available at these locations:
		- http://localhost:8080/swagger-ui.html
		- http://localhost:8080/backups
		- http://localhost:8080/exports/{backup id}