Fast, without losing messages
How Felix answers a fsynced publish in a third of a millisecond, writes at the disk's full rate under load, and slows publishers down instead of dropping their messages.
Even overall, lopsided per stream
Felix placement kept every broker's leader count near the cluster-wide share, and a single stream could still land 1/1/10 with all its writes on one broker; this is how placement now spreads each stream's leaders too.
How we catch a 1% regression
Felix's rule is that a release must not be measurably slower than the one before it; this is how same-runner A/B runs, a Welch's t-test and throwaway bisect PRs in GitHub Actions find a 1-2% loss, and why a single benchmark run can't.
One slow subscriber shouldn't slow anyone else
How Felix keeps one slow subscriber from holding up publishers or other readers, and how a 2% single-publisher loss came from where the feeder gave back a subscriber's byte budget.
Ten parcels, one envelope: Linux UDP GSO and Felix's MTU cap
Raising Felix's QUIC MTU on Linux stalled delivery for good, because UDP GSO turns ten packets into one datagram the kernel refuses and quinn never falls back; this is how we found it, why the cap is 6,550, and why 4,096 is the default.