I am now back from holiday, but I do have 100s of emails to work through and for added excitement while I was away a rat has come through my waste pipe in my kitchen and destroyed various bits of plumbing under my sink which is a nice return. So as always I'm going to be working as hard as I can but also repairing various plumbing over the next day or two, so this issue may take some time to properly think about.
It's extremely difficult for me to track these sort of issues without any concrete information as my server gets a very high number of requests constantly. Generally speaking I believe that if the server is returning ok then the request has been made sucessfully to Firebase Cloud Messaging by my server and the failure in this case is likely to be between Firebase Cloud Messaging and the app receiving the notification. This will be particularly difficuly to debug unless I can create this scenario myself.
It would be helpful if one or more users who are seeing this flakeyness can keep detailed records of any failures (times etc) and send me a bug report specifying your unique webhook id, how you are calling the WebHook (e.g. GET or POST, any body and what mechanism you call the API from). This way I can try and find what I can about an individual failure.
It's extremely difficult for me to track these sort of issues without any concrete information as my server gets a very high number of requests constantly. Generally speaking I believe that if the server is returning ok then the request has been made sucessfully to Firebase Cloud Messaging by my server and the failure in this case is likely to be between Firebase Cloud Messaging and the app receiving the notification. This will be particularly difficuly to debug unless I can create this scenario myself.
It would be helpful if one or more users who are seeing this flakeyness can keep detailed records of any failures (times etc) and send me a bug report specifying your unique webhook id, how you are calling the WebHook (e.g. GET or POST, any body and what mechanism you call the API from). This way I can try and find what I can about an individual failure.