From 1f2ac34bcdd91fc95758fd3f8672f54054213d62 Mon Sep 17 00:00:00 2001 From: zzstoatzz Date: Thu, 13 Aug 2026 20:23:52 -0500 Subject: [PATCH] validator: correct the incident comment to match the revised analysis MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The "~118s/check was queueing over a ~4s network" reading was a one-window sampling artifact; wedge-long averages show completing checks at ~0.1s. The metric split stands on its own — a conflated timer can't discriminate network from pool queueing — but the comment should not assert the retracted numbers. Co-Authored-By: Claude Fable 5 --- src/internal/validator.zig | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/src/internal/validator.zig b/src/internal/validator.zig index 348ba8f..e8490ca 100644 --- a/src/internal/validator.zig +++ b/src/internal/validator.zig @@ -865,10 +865,10 @@ pub const Validator = struct { const idx = self.acquireHostResolver(); defer self.releaseHostResolver(idx); - // pool wait and network time are separate metrics: during the - // 2026-08-13 stall the combined timer read ~118s/check when the - // network itself was ~4s — queueing behind the 4-slot pool was the - // rest, and the conflation misdirected the diagnosis. + // pool wait and network time are separate metrics: the combined + // timer misdirected the 2026-08-13 stall diagnosis, because a + // reading in seconds cannot say whether the network or the 4-slot + // pool queue is the bottleneck. const network_t0 = microTimestamp(self.io); _ = self.stats.host_authority_pool_wait_us.fetchAdd(@intCast(@max(0, network_t0 - wait_t0)), .monotonic); defer { -- 2.51.2