fix(trimmer): reduce max context token utilisation to 70%
What does this merge request do and why?
This MR decreases the context window utilization factor from 0.9 to 0.7 to provide additional room and avoid hitting context limits during conversation history trimming. This MR is needed because we are having errors in production with "prompt is too long" messages, hinting that there isn't enough buffer between what the user can use and the maximum capacity
How to set up and validate locally
Numbered steps to set up and validate the change are strongly suggested.
Merge request checklist
- Tests added for new functionality. If not, please raise an issue to follow up.
- Documentation added/updated, if needed.
- If this change requires executor implementation: verified that issues/MRs exist for both Go executor and Node executor or confirmed that changes are backward-compatible and don't break existing executor functionality.
Edited by Fabrizio J. Piva