Integration Approaches

Sugar platform is extremely powerful and easy to integrate with. Determining the right approach for your use case scenario is the key to success. 

• Push/Pull from Sugar via REST API
  
  • Recommended approach
  • Developers have full control over integration
  • Integration can be stopped and systems will continue to work
  • Retry mechanism: Resume integrations at any time (bringing only delta changes through "last_udpdate_time" flags)
  • Time delay from/to Sugar is minimum and bearable
  • Sugar REST API offers everything you need, but can be extended
• Push to Sugar, Sugar pushes to external queue in realtime
  • Uses Sugar Events to trigger through Logic Hooks
  • Requires Logic Hooks to be created and deployed
  • Does not offer "retry mechanism", if a logic hook fails to deliver to a queue or contact external system
  • Synchronous in nature, requires a response from external system
  • Needs to handle dependencies manually (contact can only be created after account)
  • Good to have a middleware/queue in front of Sugar to receive those event triggers (not mandatory but would make it "asynchronous" 
• Realtime: Embed your app into Sugar through Dashlets
  • Simplest approach
  • Often requires user-based authentication
  • All load/traffic goes directly to your system, careful with bottlenecks

REST API Webservices

Accessing your data when you want is good.  Accessing your data however you want is better. When you decide to build an integration with Sugar, you’ll need an easy way to access and interact with the data stored in Sugar. The REST (representational state transfer) API (application programming interface) is perfect for this.

• Sugar REST API Overview
  • Fully RESTfull (GET, POST, PUT, DELETE)
  • OAuth2 token based
  • All fields and modules are available through the API automatically
  • Metadata API provides everything you need to know about your Sugar instance.

• Topics
  • Intro to the Sugar REST API
  • How Sugar REST API versioning works
  • RestAPI Developer Guide
  • School of REST blog series
  • REST APIs Endpoints
  • How to add your own custom endpoints to the REST API
  • API Exceptions
  • Best practices when integrating and migrating Sugar

Common APIs

It is important to understand our most commonly used APIs and their functions within Sugar. Your integration might need data within Sugar's module structure that could potentially change over time, so you can use Metadata for it as well as CRUD onto Sugar.

• Metadata
  • Endpoint with current system's structure
  • Modules, fields, relationships, layouts 
  • Use carefully, its payload can be MBs in size, cache as much as possible
  • GET /metadata (use filters as much as possible)
  • It does not retrieve lables for dropdowns, use language api to do so
  • 3 Tips for using the Sugar Metadata API
• Language
  • Endpoint to retrieve labels
  • Dropdown list values
  • Any language specific content
  • GET /lang/<language_code>
• CRUD (Create, Read, Update, Delete) records
  • Read: GET /<module>/:record
  • Create: POST /<module>
  • Update: PUT /<module>/:record
  • Delete: DELETE /<module>/:record
  • Filter: POST /<module>/filter
    • You can use GET as well
    • POST solves the issue when a query is too long and hits HTTP limits
  • Retrieve records related to another module: GET /<module>/:record/link/:link_name
  • Creates a relationship to a pre-existing record: POST /<module>/:record/link/:link_name/:remote_id

Performance and Best Practices

• Use Sugar's sync_key
  • Sugar provides in every module a sync_key field
  • sync_key is a unique identifier for that module that links back to your external system's generated key
  • For example, account ID from ERP is AC0001, set sync_key (pass as parameter in the POST) and sugar will create a record in account with sync_key=AC0001
  • Next patch/update, sugar will check "where sync_key='AC0001'" and update it (it is indexed)
• Use Sugar Integrate Upsert APIs
  • Sugar's "update or insert" API based on sync_key
  • Replaces commonly used check if exists, update or create (GET and POST or PUT)
  • Upsert by sync_key
  • Get by sync_key
  • Delete by sync_key
  • Set a sync on existing record (without sync_key)
  • Create relationship based on sync_key
  • Delete relationship based on sync_key
• Use Sugar Bulk API
  • Endpoint that encapsulates multiple API calls into one single HTTP request
  • Requests are executed synchronously (one after the other)
  • Reducing network latency between calls
  • Combined with Upsert provides the best performance
  • Careful not to add too many requests to it
  • Start small (5 to 10 requests) and increase until you reach it's max capacity
• Polling and filtering time-based with Filter API
  • Integration gateway (or your external system) asks Sugar every X minutes
  • Request Filter API using max_num (max number of records), offset and fields
  • Response will provide the data as well as next_offset (offset can be seen as pagination)
  • if next_offset is -1, no more records to go through
• Initial Data Load into Sugar
  
  • Use Bulk + Upsert to achieve data loads
  • Millions of records may take a little while to process, plan ahead (over the weekend)
  • Use multi-thread processes to load (for example 20 threads sending 20 bulk upsert requests)
  • Careful with Sugar Cloud rate limits
  • It's always a good idea to let Sugar Support know you're data loading, they will be considerate.
• Understanding PHP limitations
  • PHP allocates OS resources to process your request
  • If you send too much data using bulk, you may run out of memory for that request
  • Push to the limits but give some room for PHP to finish the request gracefully 

You're up and running, but don't stop now - this is just the beginning of your journey. Now it's time to Customizations.