In a simple latency benchmark for memory allocation, ShenandoahPacer contributes quite a lot to the long tail latency > 10ms, when there are multi mutator threads failed at fast path to claim budget here https://github.com/openjdk/jdk/blob/fdc16a373459cb2311316448c765b1bee5c73694/src/hotspot/share/gc/shenandoah/shenandoahPacer.cpp#L230, all of them will forcefully claim and them wait for up to 10ms https://github.com/openjdk/jdk/blob/fdc16a373459cb2311316448c765b1bee5c73694/src/hotspot/share/gc/shenandoah/shenandoahPacer.cpp#L239-L277

In the test with instrumentation of the time spend on each step, shenandoahPacer often contributes multiple milliseconds, we could optimize current Shenandoah Pacer for less impact on long tail latencies.