Forum Discussion
Scheduling a Microsoft Fabric Pipeline with Specific Time Intervals
- 4 months ago
Hello JibinSebastian
I am assuming your Fabric capacity is paused outside business hours (7AM - 7PM) to save costs. If you set your pipeline to trigger every 10 mins, the trigger will run fine but your pipeline will fail because your capacity is found Paused.
Secondly, Power BI semantic model enforces a 30 min minimum interval window, so the service backend will block it, regardless we tried a 10 min interval. Maximum 8 scheduled refreshes per day per dataset is allowed.
Considering the above, I suggest you design it the following way:
Orchestrator pipeline
Notebook activity (checks against weekends)
Scheduled every 10 minutes, business hours onlyPipeline 1 – Data ingestion (trigger if parent Notebook passes as True)
Copy DataPipeline 2 – Semantic model refresh (runs after pipeline 1 success, engine may skip or throttle if run within 30 mins)
Semantic model refresh activitySchedule as discussed in previous post:
- Frequency: every 10 minutes
- Time window: 07:10–18:50
- Time zone: EST
Hi @debroshi, thank you for your response.
I currently have two pipelines:
- The first pipeline uses Copy Data to move a small amount of data from a large SQL Server table.
- The second pipeline is for refreshing a semantic model.
I’m new to Microsoft Fabric and ran into a couple of issues:
- I tried scheduling the pipeline to run every 10 minutes. It works during active hours, but when Fabric is suspended (outside 7 AM to 7 PM), the pipeline fails to execute.
- For the semantic model refresh, the scheduler only allows a minimum interval of 30 minutes. However, I need it to refresh every 10 minutes on weekdays between 7 AM and 7 PM.
Could you please guide me on how to set up the trigger properly? It would really help if you could explain it in simple, beginner-friendly terms.
Thank you!
For the semantic model refresh, the scheduler only allows a minimum interval of 30 minutes. However, I need it to refresh every 10 minutes on weekdays between 7 AM and 7 PM.
That is way too frequent. Consider using Direct Query connection or streaming datasets/RTI.
Note: If you request a refresh via API (for example through Power Automate) there is no such limit - but you must implement your own polling to find out if there is currently a refresh in progress before you submit another request.