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
While deploying the autopush-endpoint to canary we saw an usually high number of No app ID found for user errors. It's not a 100% failure rate, however, as about 50% of the traffic goes through without a hitch.
We should record the "invalid" app-id as part of the the sentry error to determine if this is a string formatting error or some other bit of strangeness.
The text was updated successfully, but these errors were encountered:
This patch does two things in order to determine the spike in `InvalidAppId` messages:
1) Add the sent app_id to the InvalidAppId error message
2) strip quotes from the app_id (in case this is a json string
transposition error.)
Closes#302
* bug: add app_id to error message
This patch addresses the spike in `InvalidAppId` messages:
* Add the sent app_id to the InvalidAppId error message
* Add minimal handling of GCM data sends. This needs to be tested on stage
(if possible. It's no longer possible to create GCM tokens to validate).
* Realized that autoendpoint was not reading credentials from files.
This is different behavior from the python version and SRE would
appreciate the option.
Note that this overloads "FCM" settings. See
router::fcm::settings::FcmSettings for details.
Closes #CONSVC-1765
Closes#302
While deploying the autopush-endpoint to canary we saw an usually high number of
No app ID found for user
errors. It's not a 100% failure rate, however, as about 50% of the traffic goes through without a hitch.We should record the "invalid" app-id as part of the the sentry error to determine if this is a string formatting error or some other bit of strangeness.
The text was updated successfully, but these errors were encountered: