I don’t think I agree that this is no longer needed. It’s great if the new updater has higher throughput, but presumably it’s not infinite, and if the edit rate is high enough that the new updater can’t keep up, then we probably want to start throttling edits, rather than have an increasingly outdated query service?

I don't have much opinion on this so I'll try to ponder the pros & cons:

Arguments in favor of removing it:

• can be cumbersome to operate (maint script running via a cron on mwmaint1002)
• could cause frustration because only well-behaved bots are actually throttled
• if we never had such lag issues on WDQS this feature would not have seen light

Arguments in favor of keeping it:

• Exactly what Lucas said, WDQS might become again the bottleneck and we don't want it to be increasingly outdated.

Overall if we decide to keep this system in place I would suggest to make sure it's in line with the objectives (lag < 10m) and adjust its scaling factor accordingly (e.g. if most bots use maxLag=5 make sure that it only breaks if wdqs lag is > 10m).

I agree with Lucas. I think we want to have a safeguard in place in case things go wild again and people are starting to push the limits even of the new updater.