M10.9: raise the evo-server session cap to 2000 and close 10.9 for now
Foundation / checks (deploy) (push) Successful in 3m26s
Foundation / checks (ecs) (push) Successful in 1m55s
Foundation / checks (client) (push) Successful in 12m32s
Foundation / checks (network) (push) Successful in 13m34s
Foundation / checks (tsan) (push) Successful in 7m42s
Foundation / checks (checks) (push) Successful in 32m43s
Foundation / checks (postgres) (push) Successful in 46m18s
Foundation / postgres-tools (push) Successful in 17s
Foundation / image (push) Successful in 2m28s
Foundation / server-image (push) Successful in 4m5s

Maintainer decision (2026-10-09): the region session cap becomes a safety bound of 2000
(plus 64 connections of headroom); the load governor is the working limit. Library bounds
go from 1024 to 4096 in SessionControl, IntentAdmission and ReplicationServer (defaults
stay 32). Drills stop for now: saturation and native-client observation move to a later
2k drill, whose prerequisites ADR-0097 lists (single-thread CPU, region login rate,
256 MiB pod limit, linear lookups). Declared: >= 32 sessions within every budget on the
reference worker. Budgets v2, STATUS, ROADMAP, VALIDATION. Refresh the dev session window.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
dingo
2026-10-09 08:22:04 -04:00
co-authored by Claude Opus 5.5
parent 24a2d4baaf
commit 1daa62cf6d
13 changed files with 91 additions and 8 deletions
+2 -1
View File
@@ -551,7 +551,8 @@ drain lease loss with 24 bots (all back within 5 s of the successor) done; no st
raw-UDP sender (abuse stays in-process/loopback; datagram-cap eviction cost analysed).
10.9d drill harness (load-drill.sh, server load-report, bundle-mode evo-load, Mbed TLS
crypto lock) done; reference runs 3-5 done (budgets v2, watermark coalescing, batched durable
lookups; within budget to the 32-session cap); saturation and native-client observation open.
lookups; within budget to the 32-session cap). Session cap raised to 2000 (maintainer, 2026-10-09); 10.9
closed for now, with saturation and native-client observation deferred to a later 2k drill (ADR-0097).
**Entry:** M9 client/session evidence and M7/M8 dev/stage durability. Read
NETWORK_ARCHITECTURE, PERSISTENCE_ARCHITECTURE and ADR-0017. The M9 transport is
+6 -1
View File
@@ -141,7 +141,12 @@ Reference runs on evolution-dev:
reached the configured 32-session cap without saturating. At 20, the login ramp drew 4
refusals and per-session p99 was 0.554 ms (budget 0.5).
Open before 10.9 completes: saturation beyond the cap, and native-client observation.
The maintainer closed 10.9 for now (2026-10-09):
- evo-server's session cap is raised to 2000, with the load governor as the working limit;
- drills stop for now. Saturation and native-client observation move to a later 2k drill.
ADR-0097 lists what that drill must solve first: single-thread CPU (the governor would
cap near 50–70 sessions), the region login rate (64 per 60 s), the 256 MiB pod limit and
linear session lookups.
## Previous gate - 2026-10-02
@@ -460,3 +460,32 @@ Open for 10.9's completion criteria, a maintainer decision:
Follow-ups:
- login and close store work off the simulation thread, or batched;
- the commit off the tick (already recorded).
**Maintainer decisions (2026-10-09), closing 10.9 for now:**
1. **Session cap raised to 2000** in evo-server (`m10_max_sessions`, plus 64 connections of
headroom). The library bounds go from 1024 to 4096 (`SessionControl`,
`IntentAdmission`, `ReplicationServer`); the library defaults stay 32. The cap is
now a safety bound, and the load governor (refuse, then shed) is the working limit.
2. **No more drills for now.** Saturation and native-client observation move to a later
2k drill. The maintainer will watch that drill with the native client.
Declared, on the evolution-dev reference worker:
- **Sustainable:** at least 32 sessions within every budget (fifth run). That is a
measured lower bound, not a saturation point.
- **Overload response:** measured as designed. The governor refused logins during a ramp
burst and shed during a mass close, and no tick debt built up.
What the 2k drill has to face (estimates, not measurements):
- **Per-tick CPU.** The session layer costs 0.35–0.55 ms per session per tick on one
thread against a 25 ms p95 budget. The governor would refuse somewhere around 50–70
sessions, well short of 2000. 2000 sessions in one region need a different architecture
(session work off the simulation thread, several regions per world, or both). That
calls for its own ADR before the drill.
- **Login rate.** `ControlConfig::max_region_attempts` is 64 per 60 s region-wide, so a
2000-session ramp would take about 31 minutes. Whether to raise it is an abuse-budget
decision for that drill.
- **Memory.** The evo-region pod is limited to 256 MiB, and the governor bounds CPU, not
memory. Each session's baselines, msquic state and per-session queues need measuring
before thousands of sessions are admitted.
- **Linear scans.** Sessions and actions are found by linear search; that is fine at tens
of sessions, not at thousands.
+21
View File
@@ -5291,3 +5291,24 @@ Image `c597c0ba…`, 11:28–11:33 UTC, restored Ready. Steps 20/24/28/32, 60 s
Not done:
- saturation beyond the 32-session cap;
- native-client observation during a drill.
## 2026-10-09 — M10.9 session cap 2000 (maintainer decision)
evo-server's session cap is 2000 (plus 64 connections of headroom). The library bounds
are 4096 in `SessionControl`, `IntentAdmission` and `ReplicationServer`. New tests check
that each accepts 2000 and refuses 4097. No drill was run, by decision: saturation and
native-client observation are deferred to a later 2k drill. Its prerequisites are in
ADR-0097.
Executed:
- WSL Release `quic`: the network test passed (after a test fix: the helper config's
`max_accounts` of 8 was below the 2000 cap), and quic/ops-contracts/session-fuzz
passed.
- The 18 Postgres suites passed; every evo-server process starts with the 2000 cap, and
the `evo-load` smoke passed.
- clang-tidy 19, clean.
- Windows `debug` 60/60 and `quic` 5/5.
Not run: Windows `client-network`. The maintainer's `evo-client.exe` was running from
that build directory, so the link could not replace it (LNK1104). The change touches no
client code.