"A second operation was started on this context instance…" under simulated latency conditions

[marjanoh2]

Description:
While testing the application with simulated latency using Clumsy (to introduce artificial delays between the app and the database), I frequently encounter the following EF Core exception:

A second operation was started on this context instance before a previous operation completed.

This appears to occur when multiple asynchronous operations are triggered close together, resulting in overlapping usage of the same scoped DbContext.
The error can also occur during rapid component navigation, when an unfinished async database operation is still in progress while a new one begins, both sharing the same scoped DbContext instance.

Question:
Would using IDbContextFactory or DbContextFactory in the handlers be a better and safer long-term approach to prevent this issue—especially under high-latency or concurrent query scenarios?

[neozhu]

Yes, I remember fixing this issue before. But if a page runs multiple database queries at the same time, the error can still happen.

[marjanoh2]

The issue most frequently occurs during navigation between pages or components, especially when switching quickly. It seems that some async operations involving the DbContext are still running when a new component starts its own database call, resulting in overlapping usage and the “A second operation…” error.

[neozhu]

Got it. I’ve found a solution to this issue, but it requires a relatively big change. The key is to create a new DbContext instance in each handler using IDbContextFactory. This approach avoids shared context issues and prevents the error, especially under concurrent or delayed operations.

[marjanoh2]

This change is well worth it for stability. The error may not happen often, but when it does, it affects reliability. DbContextFactory ensures safe and isolated usage in all cases.