You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Mar 22, 2023. It is now read-only.
When computing fallback-period-secs and stale-timeout-mins for a service that has multiple source tokens, the current implementation just grabs these values from the last token in the list that sets them. The fallback logic should be more robust, e.g., taking the maximum value.
Alternatively, simply dropping support for combining tokens like this would also solve the problem. The biggest reason we haven't actually seen this bug in production is that no one uses this feature now that we support parameterized services.
The text was updated successfully, but these errors were encountered:
When computing
fallback-period-secs
andstale-timeout-mins
for a service that has multiple source tokens, the current implementation just grabs these values from the last token in the list that sets them. The fallback logic should be more robust, e.g., taking the maximum value.waiter/waiter/src/waiter/service_description.clj
Lines 1230 to 1234 in bc04a0b
Alternatively, simply dropping support for combining tokens like this would also solve the problem. The biggest reason we haven't actually seen this bug in production is that no one uses this feature now that we support parameterized services.
The text was updated successfully, but these errors were encountered: