Nebol's Wiki

NebolIntegration is an active project with on-going development, and the documentation is currently lacking. Chunks of it still needs to be written. I will however answer questions to the best of my abilities.

NebolIntegration is a light-weight C# library that manages integrations between systems.

See the nuget page at https://nuget.nebol.se for version history etc.

NebolIntegration requires a number of tables to be created in the storage of your choice. They are prefixed with “Integration”. You will create a class containing the actual integration logic. This class is then provided to the integration engine that will handle the rest.

• Handles incoming and outgoing integrations
• Synchronous and asynchronous integrations
• Extensive tracking, logging and error-handling
• Failed tasks are tried again according to specified schedule
• Tasks can easily be re-sent, re-processed, and responses can be re-processed
• Healthchecks, built-in or custom coded, with configurable intervals and thresholds
• Statistics
• Optional triggers using crontab-format to process integrations
• Scheduling for special handling during maintenance windows etc
• Reports and notifications
• Automated maintenance (log cleaning)
• Grouping of integration processes simplifies administration
• Optional throttling for demanding processes

NebolIntegrationNotifier (windows application, downloaded separately) provides

• Notification by email when an integration task fails
• Notification when a healthcheck fails (or when it is returned to healthy state)
• Daily reports of all integrations with statistics per day, week, month, year

• Data translation

Storage for the integration data such as an Entity Framework supported database or any other storage supported by SharpRepository (MongoDb, RavenDb, Db4o, XML etc)

The functionality of NebolIntegration is rather straight-forward but still there are a bunch of concepts to understand in order to properly implement an integration. I think that it's perhaps best explained by a couple of tutorials…

Example 1 - Outgoing asynchronous integration using sqllocaldb and GUID generating webservice

NebolIntegration_Basic_Integration - Creating your first integration

NebolIntegration_Database_and_Flow - So what happened?

NebolIntegration_HealthChecks - What are HealthChecks? Adding a healthcheck.

NebolIntegration_Triggers - What are triggers? Adding a trigger.

NebolIntegration_RetryScheme - How the retry scheme works

NebolIntegration_Incoming_Integration - Creating an incoming integration

NebolIntegration_Reference_and_Message - Managing user data (also static vs dynamic tasks)

NebolIntegration_More_Integration_Variations - Synchronous integrations, overrides, various trickery

NebolIntegration_Builtin_Integrations - Updating, Cleaning, Statistics, Reports

NebolIntegration_Maintenance - Into the database we go (changing status and overriding with commands)

NebolIntegration_Notifier - Setting up and using the NebolIntegrationNotifier

NebolIntegration_Console_Commands - Utilising the built-in console commands

Setting up database

Configure the DefaultConnection and run this in Package Manager:

update-database -ConfigurationTypeName NebolIntegrationDbContextConfiguration -Verbose

This will create the necessary tables

The default Configuration-class looks like this:

How to specify other connection string than DefaultConnection ??

Setting up SharpRepositiry

Running the Integration

Additional topics: The database (manual configuration, troubleshooting integration tasks)

Some breaking changes have been made since the early versions:

SystemIntegration renamed to Integration

Instead of a ISharpRepositoryConfiguration instance the Engine, Integrations etc are now accepting a NebolRepositoryConfiguration instance which has a ISharpRepositoryConfiguration instance plus an optional repository name

Version 1.1.x.x will support “Triggers”, that will trigger processing automatically according to a set cron-schedule thus no Submit will need to be made in the ValidatePolling() method (which was one way to go about it). Also full support for the built-in integrations “Updating” and “Cleaning”

Version 1.2.x.x will support running multiple engines (clients) & multiple threads per integration.

Version 1.3.x.x will support Schedules that you can setup to plan ahead for maintenance windows, or if you only want the integration to be run on business hours etc.

Still not fully planned:

Built-in integrations “Statistics” and “Reports”

The additional tables Machine & Application needed for multiple client support

Notifying clients (via built-in method or RabbitMQ) instead of client polling

Improved built-in UI for command console using NebolConsole and multiple console areas

Console command for listing & viewing queue & result items

Web UI for administration + links in report-mail for errors etc

fixing the issue of direction not being read properly from the db