Details
-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Fix Version/s: None
-
Component/s: Design Documents
-
Labels:
-
Story Points:2
-
Epic Link:
-
Team:System Management
Description
LDM-151 section 3.1.2 talks about an "OCS Database" that is provided to the TCS, either by running SFM at the summit or by persisting to a database at the "data center" that is accessible by the TCS.
First, the scheduler is part of the OCS, not the TCS. Second, DM has no plans to run SFM at the summit (although there have been discussions about possibly running it within the Camera Diagnostic Cluster). Third, no database at NCSA, where the Alert Production runs, can be directly accessible by the TCS nor OCS.
Instead, the plan is for the relevant quality metrics to be provided to the scheduler component of the OCS via telemetry that begins either at the Alert Production pipeline itself or at the Prompt Quality Control system, proceeds via a DM-internal messaging system (likely RabbitMQ) to the Telemetry Gateway at the Base, is converted into telemetry messages for the OCS SAL system, and then is communicated to both the scheduler and the EFD, which plays the role of the "OCS Database".
This section of LDM-151 should be rewritten to either reflect this or delegate the topic to other design documents.
K-T — I've made a minimal change to address this at https://github.com/lsst/LDM-151/pull/60. I don't think it's appropriate for LDM-151 to get into the details of how this is implemented, and I'm not aware of any other relevant design documentation beyond a sentence in LDM-148, but I'd be happy to add further references if you can suggest some.