Investigate end-to-end pipeline latency
Our pipeline currently has an average latency of ~9s. This is in line with what Chichi found in her paper: https://journals.aps.org/prd/pdf/10.1103/PhysRevD.105.024023
(Table II, page 12)
There's some possible improvements discussed in that paper, but we can also start our own investigation.
We want to:
- Reproduce Chichi's results
- Attempt the suggested improvements
- Monitor the pipeline's latency on a more granular level
- (Ideally through Grafana)
- Consider other latency improvements.
These could each be separate issues. If we find a larger project, this could spin off into a milestone (or epic).