The new system has been running in production without any major issues for a few weeks now. In a few weeks, I plan to start using the second collector, and then announce the new system to Project members to tell them how they can use its new features.
We moved forward with the implementation, and the new job queue system is now being tested in production on a single test pull request. Most things seem to be working, but there are a few things to iron out and some profiling to be done. I expect that within a few weeks we could be ready to switch to the new system fully in production.
It is possible to now run the system with two different machines on two different architectures however there is work to be done to make this more robust.
We have worked on ironing out the last bits and pieces for dequeuing benchmarks as well as creating a new user interface to reflect multiple collectors doing work. Presently work is mostly on polishing the UI and handing edge cases through manual testing.
