Textline incident
Inbound Message Delays in Conversation Windows
Textline experienced a minor incident on October 3, 2025 affecting Textline Web Application, lasting 2d 22h. The incident has been resolved; the full update timeline is below.
Affected components
Update timeline
- investigating Oct 03, 2025, 07:06 PM UTC
We are investigating a spike in our message sync queue beginning around 10:00 AM PT. This is causing intermittent delays where inbound messages appear in the conversation list but do not immediately display in the conversation window. Conversations eventually sync, but with noticeable lag. We are actively monitoring error logs and working to restore normal performance.
- identified Oct 03, 2025, 10:22 PM UTC
The issue has been identified and we are working to apply a fix.
- resolved Oct 06, 2025, 05:26 PM UTC
This incident has been resolved; inbound messages are populating into the application appropriately and are no longer experiencing any sincere delays in doing so.
- postmortem Oct 09, 2025, 07:25 PM UTC
Root cause: Significant outbound messaging volume created downstream impact for the primary application writer database, as it had difficulty updating message credit pools at the rate that messages were being sent outbound. Impact: Application latency increased outside of normal bounds, causing minor but widespread delays in application interaction. A small number of accounts experienced more acute errors, primarily falling into two categories. Error messages: Local user caches fell out of sync with the application, causing various errors. Clearing the local cache resolved these issues. Message synchronization: One of 60 background workers froze due to its interaction with the database. Inbound messages were received by Textline's proxy application, but not properly synchronized into the primary application database where it can be viewed by customers. Synchronizing the stuck messages is still under investigation. Primary Resolution: Database load related to message credit pool updates has been significantly decreased by isolating credit decrement functionality from real-time database updates. Database sizing has also been significantly increased as an extra precaution.