RT visibility receive signal display developed and integrated into the SDP prototype

Once SDP is deployed as part of a prototype integration system or early array release, operators will require a mechanism to easily assess visibility data being received by the SDP in order to take action if appropriate.

For this it is assumed that a simple signal display capable of plotting visibility amp/phase vs channel for either all or selected baseline(s) will be required with a view that this can be extended to additional plots once an initial system is integrated and demonstrated.

<To be discussed / agreed with the FO at PI planning!>

• Simple working PoC for real-time visibility signal display consisting of line plot of amp/phase vs channel for selected baseline(s)
• Demonstration of this using the updated SDP visibility receive workflow and/or appropriate alternative test workflow.

Potential stretch goals:

• Integration into SKAMPI including how an operator might be able to discover and access the plots / display when the prototype is deployed on an integration cluster (work with the system or platform team on this!)
• Consideration for how signal display plots might be selected / specified / configured
• Extension to other plots / displays in addition to the simple line plots.

The Python metric generator is available in GitLab at https://gitlab.com/ska-telescope/cbf-sdp-emulator-metrics-generator
It can be run either stand-alone using the GLEAM datasets (eg GLEAM large with four channels) or else integrated with the CBF-SDP Emulator (https://gitlab.com/ska-telescope/cbf-sdp-emulator), where it works as a consumer.
It supports multiple metrics, where each metric can independently filter selected specified polarisations and baselines, aggregage across any number of specified channel intervals, aggregate across specified number of time intervals. It operations on real, imaginary, amplitude or phase for the visibility data and prodices either mean, max, min, or variance for aggregated data each time block.
It currently pushes data via HTTP POST requests to InfluxDB, and Grafana dashboards are used for visualisation.
With some concerns about the performance of the Python metric generation, NZAPP has also enhanced the RDMA receiver code which can stress-test the InfluxDB-Grafana combination.
This work is presented in System Demo 7.5, with the attached slides.