Data streaming: run starts: standardise on JSON schema for metadata structure

[rerpha]

https://github.com/ess-dmsc/streaming-data-types/blob/master/schemas/pl72_run_start.fbs#L29 allows for metadata such as the user names and other data which we will need to put in the run start message.

This ticket is to document these so that the keys remain the same between whatever produces them( likely kdae_control) and any downstream consumers such as the journal appender started in #11

acceptance criteria

• A JSON schema is added to this repo that sets the key names of basic fields we would like to the metadata. These should be the ones in the journal XML files.

---

Discussed Planning 08/01/2026 14:36

[Tom-Willemsen]

This may take inspiration from https://github.com/bluesky/event-model which is also a set of JSON schema definitions for how to serialise and deserialise JSON documents with specific keys and meanings.

[Tom-Willemsen]

Should include time-channels for use by downstream consumers (even if hardware doesn't use them)

[Tom-Willemsen]

Putting all the metadata in the runstart message means that it cannot be edited after the run has started. Is that acceptable for HRPD-X?

The ICP lets you edit most metadata things like users, rb number, title up until run end.

[rerpha]

which metadata would a user typically change?

[Tom-Willemsen]

Users, rb number, run title were the ones I was thinking of.

[rerpha]

ok - will ask the scientists but i think it would be OK to not permit changing those during the run, making it clear they need to be set beforehand.

[rerpha]

This will only be metadata for kafka_dae_diagnostics - the filewriter will use static fields in the nexus_structure

curently kdaediag would like to know about:

• num_periods
• num_time_channels
• time regimes
• vetos