jinghui/hermes-webui
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
6,531 Commits 1 Branch 0 Tags
af67bd5968be33479b60c193593431fdb2e490ad
Commit Graph
3 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
Konstantin M af67bd5968 perf(webui): Priority 1+4 - cheap user_message_count + profiles TTL bump 
Priority 1: Session.compact() walked all self.messages to count user-role
messages. For a 2,730-msg session that's ~1.5s on eMMC. Replaced with
Session._compute_user_message_count_lazy(): single indexed SQLite query,
~5ms, same correctness. Field is consumed by sidebar-row code
(_looks_like_stale_zero_message_row, _row_may_need_sidecar_metadata_refresh)
so we MUST keep emitting a real value, not None.

Priority 4: _LIST_PROFILES_CACHE_TTL 4s -> 60s. Invalidation hooks
(create_profile_api, delete_profile_api) already call
_invalidate_list_profiles_cache() so the bump is safe.

Bench (10 iters, 10s timeout, live Chromebook eMMC):
  long_session_idle   p50  5759ms -> 2869ms  (-50%)
                     p95  9141ms -> 3156ms  (-65%)
                     max  9740ms -> 3283ms  (-66%)
                     timeouts 2/10 -> 0/10
  session_switch_idle p50  2336ms -> 1316ms  (-44%)
                     p95  3708ms -> 1442ms  (-61%)
                     max  4218ms -> 1525ms  (-64%)
  profiles_idle       p50    6ms ->    6ms
                     p95  821ms ->  248ms   (-70%)
                     max 1488ms ->  445ms   (-70%)
  sidebar_idle        p50  123ms ->   31ms   (-75%)
                     p95  887ms ->   57ms   (-94%)
                     max  986ms ->   59ms   (-94%)
  normal_session_idle p50  359ms ->   56ms   (-84%)
                     p95  749ms ->  131ms   (-83%)
                     max  794ms ->  167ms   (-79%)

Revert: git checkout master && systemctl --user restart hermes-webui
 2026-07-06 15:40:29 +00:00
Konstantin M ce1aaf9d7b perf(webui): auto-log slow /api/session requests with stage breakdown 
Adds auto-logging when total /api/session latency exceeds 2s, so we don't
need HERMES_DEBUG_SLOW env var to diagnose regressions in production.
Env var still forces logging on every request for development.

Baseline data on the Chromebook (Celeron N3350, eMMC storage):
- long_session_idle (2,445 msgs) p50=5.8s, p95=9.1s, max=9.7s, 2/10 timeouts
- stage breakdown for a 2.8s request:
    get_session=1041ms, compact=1717ms, redact=31ms, json_write=8ms
- compact stage iterates all messages for user_message_count and runs
  recursive redact_session_data + json.dumps -- biggest single cost

Ref: perf/session-load-latency
 2026-07-06 15:21:12 +00:00
Konstantin M 1018732a26 perf(webui): Phase 0 instrumentation for session-load hot path 
The slow-request journal showed /api/profiles, /api/models, and /api/session
fire on every session click but had no per-stage timing. Add:

- RequestDiagnostics coverage for GET /api/profiles, /api/models, /api/session
- Stage markers inside /api/profiles (list / active lookup / isolated check)
- Stage markers inside /api/models (freshness routing, serialize)
- Per-stage stage-log inside get_available_models_for_session_visit that
  emits a [SLOW] line when total wall time crosses HERMES_DEBUG_SLOW ms
  (default 500ms) so we can pinpoint which sub-step is blocking

No behavior change. Existing _t0.._t6 timing on /api/session and the
2.5s/10s TTL on /api/sessions are unchanged. Reversible: revert commit.
 2026-07-06 14:23:34 +00:00