Guide to Integration Design
Purpose
The BitaBIZ API enables integration with ERP systems, HR platforms, Business Intelligence solutions, data warehouses, and custom applications.
This guide describes the recommended integration patterns to ensure integrations are efficient, stable, and place minimal load on the platform.
Basic Principles
Only fetch the data you need.
Avoid retrieving unchanged data again.
Store (cache) master data locally where relevant.
Minimize the number of API calls.
Design the integration to handle temporary errors and retries.
Initial Synchronization
During the first integration, you can perform a full synchronization of relevant data, such as employees, departments, teams, projects, job codes, and historical time registrations. A full synchronization should normally only be done once.
Ongoing Synchronization
After the initial import, the integration should only fetch new or changed data. Full synchronization should only be performed if there is a specific business need.
Employee Master Data
Master data rarely changes. We therefore recommend synchronizing once daily or only when changes are expected.
Time Registrations
Import historical time registrations once. After that, only synchronize relevant periods or changed data. It's usually not necessary to fetch many years of history every day.
Business Intelligence and Reporting
For integration with Power BI or data warehouses, we recommend a one-time historical import, local storage of the history, and then only incremental updates.
Planning Synchronizations
To ensure the best possible experience for all customers, BitaBIZ recommends scheduling synchronizations and large data exports outside of the platform's peak hours.
The platform typically experiences the highest load on weekdays between 08:00 and 10:00.
This is especially important for full synchronizations, historical data exports, Power BI updates, data warehouse synchronizations, and other large batch jobs. Business-critical integrations can, of course, be run whenever needed.
Effective Integrations
Only fetch necessary data.
Avoid repeated requests for the same data.
Avoid unnecessary historical data exports.
Only synchronize relevant periods.
Reuse already synchronized data.