When there is an exception we get from another subsystem shows that there is a mutation_write_timeout exception, but nothing found in Scylla logs while an update_table request failed. Is there any exception catch in ops like update _table, get_item.
Timeouts are not logged in general. The main reason is that their frequency can be high and they would generate a lot of noise or huge logs.
1 Like
If there is any light weight methods to log mutation_write_timeout or mutation_read_timeout in info level or error level?
You can try enabling debug-level logging for the storage_proxy logger. Note that this will probably result in a lot of log messages.