[Style] Discretization fails on large layers due to 30s database timeout

[bruhnild]

Context:
When performing a discretization (styling) on a layer containing several million entities, the SQL query executed by TerraVisu in the database fails to complete before the current timeout (30 seconds).
As a result, the process is interrupted before completion, causing an error and preventing the style from being updated.

Steps to reproduce

1. Load a layer with several million entities in TerraVisu.
2. Run a discretization through the back-office (e.g. a Jenks classification).
3. Observe that the query exceeds the 30-second timeout and fails.

Expected behavior

The discretization query should be able to complete successfully without errors, even for very large layers.

Possible solutions

Two possible improvements:

1. Increase the query timeout

   • Add a configuration variable in TerraVisu
   • Default value: 30 seconds
   • Allow it to be increased (e.g. 120 seconds) depending on the server’s capacity

2. Run the discretization as an asynchronous task (via Celery)

   • Would prevent the web server from being blocked during processing
   • Requires more development effort but would be more robust in the long term