While building Secure Notes for Jira, we paid special attention not only to security, but also to runtime performance and observability in real enterprise environments.
Initially, we integrated PostHog for external analytics. For the final submission, we decided to remove external analytics dependencies to keep the solution lightweight and fully Forge-native.
The Challenge: Performance vs. Visibility
We still wanted deep insight into database behavior. When a query exceeds a threshold (for example, >800ms), we need to capture EXPLAIN ANALYZE to understand why it was slow.
Previously, this diagnostic step was synchronous, which created a paradox: slow requests became even slower due to in-line analysis.
The Solution: Non-Blocking Diagnostics
We improved our forge-sql-orm layer to use Forge Async Events (@forge/events).
Now, when a performance threshold is breached, the diagnostic task is offloaded to a background queue.
This allows the user request to complete immediately, while SQL analysis runs asynchronously in the background.
The Result
We get full visibility into real customer query behavior and execution plans without impacting runtime performance or user experience — staying fully Forge-native and Runs on Atlassian compliant.
For more details on the original observability design: https://community.developer.atlassian.com/t/practical-sql-observability-for-forge-apps-with-forge-sql-orm/97237
Log in or sign up for Devpost to join the conversation.