diff --git a/Dockerfile b/Dockerfile
index 1b60a63..d991b0d 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -8,8 +8,9 @@ RUN curl -fSL https://ziglang.org/builds/zig-x86_64-linux-0.16.0-dev.3059+42e33d
 ENV PATH=/opt/zig-x86_64-linux-0.16.0-dev.3059+42e33db9d:$PATH
 WORKDIR /build
 
-# Uring networking patch preserved in patches/ for when upstream stabilizes
-# (see docs/evented-attempt.md). Not applied — using Io.Threaded backend.
+# patch Io.Uring networking (stubbed as Unavailable upstream, zig#31723)
+COPY patches/ patches/
+RUN patch /opt/zig-x86_64-linux-0.16.0-dev.3059+42e33db9d/lib/std/Io/Uring.zig < patches/uring-networking.patch
 
 # fetch dependencies first (cacheable — only changes when build.zig.zon changes)
 COPY build.zig build.zig.zon ./
@@ -17,7 +18,8 @@ RUN zig build --fetch-only 2>/dev/null || true
 
 # then copy source and build
 COPY src/ src/
-# ReleaseSafe: Evented GPF was the reason for ReleaseFast — back to Threaded now.
+# ReleaseSafe: the production SIGSEGV was a websocket bug, not a fiber issue.
+# trying Evented + ReleaseSafe — if the repro GPF hits production, fall back to ReleaseFast.
 RUN zig build -Doptimize=ReleaseSafe -Dcpu=baseline -Dtarget=x86_64-linux-gnu
 
 FROM --platform=linux/amd64 debian:bookworm-slim
diff --git a/docs/evented-attempt.md b/docs/evented-attempt.md
index b598deb..10757bd 100644
--- a/docs/evented-attempt.md
+++ b/docs/evented-attempt.md
@@ -1,9 +1,12 @@
-# Evented backend attempt (2026-03 to 2026-04)
+# Evented backend (2026-03 to present)
 
-we spent roughly a month trying to run zlay on `Io.Evented` (io_uring fibers)
-instead of `Io.Threaded` (one OS thread per task). the goal was to drop from
-~2,800 threads to ~35. it didn't work out — the upstream fiber machinery is
-still experimental and unstable. this doc records what we did so it's not lost.
+log of getting zlay running on `Io.Evented` (io_uring fibers) instead of
+`Io.Threaded` (one OS thread per task). goal: drop from ~2,800 threads to ~35.
+
+after 28 commits fixing cross-Io issues and a brief revert to Threaded, we
+discovered the production SIGSEGV was a websocket handshake bug (TCP split
+mid-CRLF, fixed in websocket.zig `9ac64da`), not a fiber context-switch
+issue. Evented backend is back, now running ReleaseSafe.
 
 ## what we built
 
@@ -70,14 +73,16 @@ configurations.
 - `docs/stdlib-patches.md` — catalog of all 6 workarounds we needed
 - this document
 
-## when to try again
+## timeline
+
+- 2026-03-08: first Evented deploy (`39134d1`)
+- 2026-03-08 to 2026-04-04: 28 commits fixing cross-Io crashes
+- 2026-04-05: reverted to Threaded after SIGSEGV blamed on fiber machinery
+- 2026-04-05: discovered real cause — websocket handshake TCP split bug
+- 2026-04-05: fixed websocket.zig (`9ac64da`), re-enabled Evented + ReleaseSafe
 
-revisit when:
-1. upstream implements uring networking (zig#31723 closed)
-2. `fiber.zig` gets meaningful changes (context switch rewrite or fix)
-3. a zig release (not dev build) ships with Evented marked as stable
+## fallback
 
-the dual-Io architecture, DbRequestQueue, and cross-Io segregation patterns
-are all still in the codebase (dead code paths behind `Backend == Io.Evented`
-checks). switching back is a one-line change in `src/main.zig:58` plus
-re-enabling the patch in the Dockerfile.
+if Evented + ReleaseSafe hits the repro GPF in production, switch to
+ReleaseFast in the Dockerfile. if that also crashes, flip `Backend` back
+to `Io.Threaded` in `src/main.zig:58` — one-line change.
diff --git a/src/main.zig b/src/main.zig
index 5cbe314..4a8231c 100644
--- a/src/main.zig
+++ b/src/main.zig
@@ -48,18 +48,17 @@ const log = std.log.scoped(.relay);
 pub const default_stack_size = 8 * 1024 * 1024;
 
 // -- Io backend selection --
-// Io.Threaded: one OS thread per concurrent task (~2,800 subscribers + services).
-// not what we want long-term — Evented (fibers on io_uring) drops thread count
-// from ~2,800 to ~35 — but the upstream fiber machinery is not ready.
+// Io.Evented (fibers on io_uring): ~35 threads instead of ~2,800 (one per PDS).
+// Networking via patched Uring.zig (patches/uring-networking.patch) — implements
+// listen, accept, connect, read, write, send via io_uring opcodes. DNS (netLookup)
+// is NOT patched; subscribers resolve hostnames through pool_io (Threaded) instead.
 //
-// We built and shipped a full Evented backend (see docs/evented-attempt.md and
-// patches/uring-networking.patch). After 28 commits fixing cross-Io crashes,
-// heap corruption, mutex incompatibilities, and a ReleaseSafe GPF in fiber
-// context-switch, the remaining bug is a probabilistic SIGSEGV under ReleaseFast
-// (~every 30-90 min under load). Root cause is in std.Io.fiber.contextSwitch
-// (zig 0.16-dev stdlib). As of dev.3091, fiber.zig is unchanged and Uring
-// networking is still fully stubbed upstream. Switch back when upstream stabilizes.
-const Backend = Io.Threaded;
+// History: 28 commits of cross-Io fixes (see docs/evented-attempt.md). The
+// production SIGSEGV that prompted a Threaded revert turned out to be a websocket
+// handshake bug (TCP split mid-CRLF), not a fiber issue. Fixed in websocket.zig
+// 9ac64da. The repro (scripts/repro_evented.zig) still GPFs under ReleaseSafe
+// but production code paths may differ — deploying ReleaseSafe to find out.
+const Backend = Io.Evented;
 
 var backend: Backend = undefined;
 var debug_threaded_io: Io.Threaded = undefined;