Hi, team! I found the task lock granularity of index kafka task was

timeChunk

in metadata table

druid_tasklocks

. The lock of index kafka task should be segment lock. Why does the metadata table stores ``timeChunk``?

Do you have mixed segment granularity in this table? Per the docs the segment locking mode is a suggestion, not directive, so Druid may choose not to use if under certain conditions (which I am not well versed on).