Google Docs Integration, Faster Reports, and an Expanded PM Agent

Google Docs integration for documentation access. Users can now connect Google Docs as a source of internal documentation, allowing PM Agent to draw on document context when answering questions and generating insights.

Chat logic moved to a dedicated sub-agent. All chat-related logic has been extracted into a separate sub-agent. This improves stability, simplifies maintenance, and accelerates the development of communication-related features within projects.

Faster loading for My Projects. The project list loading and rendering have been optimized. The page now opens significantly faster, even for users with a large number of projects.

Heavy reports and queries are optimized. The Project Feed report on the Project Report page, along with margin and labor cost reports, has been substantially accelerated. This reduces system load and speeds up analytics delivery for all users.

Financial data added to PM Agent. PM Agent now has new capabilities for displaying and analyzing financial data, enabling decisions based on up-to-date budget and expense information.

Simultaneous import of projects and employees. Projects and employees can now be imported at the same time, speeding up initial setup and onboarding for new teams.

Fixed project import error from .xlsx files. Resolved an issue where importing projects from a correctly formatted .xlsx file triggered the error "Failed to provision user in enji-auth. Status code: 403."

New email template for Conference Bot. A modern email notification template for Conference Bot has been developed and is ready for rollout. The email includes a concise summary of the call.

Correct leave display and request filtering. Fixed an issue where an employee would briefly see all leave requests before the participant filter was applied.

Employee icons in the labor cost table are clickable again. Employee icons in the labor cost table are now clickable once more, making it easier to navigate and view colleague information.

Fixed the hour count display in the labor cost table. Hours worked now display correctly instead of showing NaN.

Resolved duplicate requests and 504 errors on the Tasks tab. Fixed an issue where the frontend was sending duplicate requests and the backend was returning 504 errors when working with the Tasks tab.

Alerting added for recurring failures in critical Airflow DAGs. The team now receives immediate notifications when key background processes fail, enabling faster incident response.

Improved "Out of memory handling" for Celery in production. Measures have been implemented to prevent crashes caused by memory exhaustion in Celery, improving overall system stability.