When using the "couchbase" output mode, documents are inserted into Couchbase with the string data format. This can be seen when querying the document as the low order bits of META(bucket).flags contains 0x4.
If instead the data is inserted as JSON data, the low order bits of META(bucket).flags would be 0x2, indicating JSON data.
When this flag is incorrect it can cause SDKs to interpret the data incorrectly. In particular, I observed it causing deserialization errors in the .Net SDK.
When using the "couchbase" output mode, documents are inserted into Couchbase with the string data format. This can be seen when querying the document as the low order bits of META(bucket).flags contains 0x4.
If instead the data is inserted as JSON data, the low order bits of META(bucket).flags would be 0x2, indicating JSON data.
When this flag is incorrect it can cause SDKs to interpret the data incorrectly. In particular, I observed it causing deserialization errors in the .Net SDK.