Different message offset (kafka) per schema(topic).

[YooSunYoung]

Beamlime should start consuming from different offset for different schema (type of data)

• PL72: Latest run start message, and match with the run_number or proposal_id.
• EV44: From the start_time of the run_start_message.
• F144: From the latest message regardless of the start time.
• TDCT: From the start_time of the run_start_message.
• DA00: From the start_time of the run_start_message.

It is because logs like detector position might be set in the past and never moved.

[SimonHeybrock]

Is this actually true? I had assumed that we would get a message in all cases when we start streaming, just like the filewriter will always get values? Only the timestamp may be in the past. Should check with ECDC before doing any work here!

[YooSunYoung]

Is this actually true? I had assumed that we would get a message in all cases when we start streaming, just like the filewriter will always get values? Only the timestamp may be in the past. Should check with ECDC before doing any work here!

It doesn't change where we set the offset of the messages though.