Make egress warnings easier to investigate after a spike #53428
Unanswered
benjaminlouwrens-gif
asked this question in
Feature Requests
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Supabase support directed me to the egress discussion, which is locked. I would like to suggest more actionable information in egress warnings and usage reports.
My clipboard sync app was polling every five seconds and downloading its full history, including images already on the device. On November 13 it made 10,570 requests. We estimated about 90 GB of responses based on the compressed history size, close to the dashboard's 90.9 GB database API spike. That response total is an estimate, not a direct measurement. The detailed logs had expired by the time I investigated, so identifying the cause took longer. I have since changed the app to check item IDs and download only missing items.
It would help if an egress warning included:
I understand API gateway logs do not currently include response-byte data. Even a service-level breakdown, request counts by route, and a clear indication of which byte measurements are available would make the warning easier to act on. Thank you for considering this and for the grace period that gave me time to investigate and fix the app.
All reactions