I have very weird issues with database on Cloud since last database incident. It happens randomly and I can't really fix it.
Straight forward example:
val documents = databases.listDocuments(...)
documents.forEach { document ->
databases.updateDocument(document.id, data)
databases.deleteDocument(document.id)
}
val document = databases.getDocument(documentId)
databases.updateDocument(document.id, data)
Yes, documents are 100 percent in database, otherwise fetching wouldn’t work, and it fails even when creating fresh document.
How could createDocument function with unique document id throw 404?
It happens to approximately 2-3% of write requests, which is not huge but it should not happen ever in the first place.
It basically corrupts users data and state since I can’t really fix or handle it properly on my end.
Any clues?
TL;DR
Issue: Developers are experiencing weird database errors on Cloud post a previous incident, resulting in random 404 errors despite documents being present in the database. Errors occur during update or delete operations, even with freshly created documents.
Solution: The unique document ID in the createDocument function may be causing the 404 errors. This issue affects 2-3% of write requests, impacting user data and states. Developers need to investigate the createDocument function and potentially modify the unique ID generation process to resolve the problem and prevent data corruption.
