Core.ScriptedAlarms-009 resolution: replace the per-call Dictionary +
AlarmPredicateContext allocation with a per-alarm reusable AlarmScratch
held in _scratchByAlarmId, refilled in place under _evalGate on each
evaluation. The hot path no longer allocates per upstream tag change.
Why this matters:
On a busy line where many tags feeding many alarms change frequently,
the old BuildReadCache allocated a fresh dictionary + context on every
predicate evaluation — a steady stream of short-lived allocations the
GC eventually has to reclaim. With the reuse, the dictionary and
context are allocated once per alarm (on first evaluation) and refilled
in place across every subsequent re-eval.
Implementation:
- New private AlarmScratch class holds the reusable
Dictionary<string, DataValueSnapshot> read cache (pre-sized to the
alarm's Inputs.Count) and the AlarmPredicateContext that wraps it by
reference. The context observes refilled values without being
re-created.
- ConcurrentDictionary<string, AlarmScratch> _scratchByAlarmId on the
engine, cleared in LoadAsync alongside _alarms so a config-publish
drops the prior generation's scratch (Inputs / Logger may change).
- EvaluatePredicateToStateAsync looks up scratch via GetOrAdd, calls
the new RefillReadCache(Dictionary, IReadOnlySet) helper to clear +
repopulate the dictionary in place, then runs the predicate against
the reused context.
- BuildReadCache removed.
Safety:
Reuse is serialised under _evalGate which guarantees no two threads
ever observe the same scratch in a half-refilled state. The
AlarmPredicateContext is bound to the scratch dictionary by reference,
so the predicate's ctx.GetTag(path) sees the freshly-refilled values
rather than a stale snapshot.
Verification:
- All 66 ScriptedAlarms tests pass (was 63 — three new regression tests
locking the reuse contract).
- All 56 VirtualTags tests still pass (unchanged).
- All 104 Core.Scripting tests still pass (unchanged).
New tests in ScriptedAlarmEngineTests:
- Reevaluation_reuses_the_same_read_cache_dictionary — asserts
ReferenceEquals(scratch_before, scratch_after) across two
evaluations of the same alarm.
- Reevaluation_reuses_the_same_predicate_context — same, for the
context.
- LoadAsync_drops_the_prior_generations_scratch — asserts a config
publish wipes the prior scratch (so a stale Logger / Inputs can't
leak into the new generation).
Internal test hooks TryGetScratchReadCacheForTest /
TryGetScratchContextForTest added via the existing
InternalsVisibleTo for the tests project. Kept internal — not part of
the public engine surface.
Docs:
- docs/v2/Galaxy.Performance.md "Scripted-alarm engine" section
rewritten as "hot-path allocation reuse" documenting the new
contract + reuse safety reasoning + the three regression tests.
- code-reviews/Core.ScriptedAlarms/findings.md -009 flipped
Won't Fix → Resolved.
- code-reviews/README.md regenerated.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- Core.ScriptedAlarms-003: emit OnEvent OUTSIDE _evalGate by collecting
pending emissions during the gate-held section and flushing them after
release; eliminates re-entrancy deadlock the docs already promised.
- Core.ScriptedAlarms-006: track every fire-and-forget Reevaluate /
ShelvingCheck task in _inFlight; Dispose drains the set so the engine
no longer races store writes against teardown.
- Core.ScriptedAlarms-008: store comments as ImmutableList<AlarmComment>
so AppendComment is O(log n) instead of O(n).
- Core.ScriptedAlarms-010: document the deliberate input-quality
asymmetry (Uncertain drives the predicate, renders {?} in the message)
in docs/ScriptedAlarms.md and on MessageTemplate.Resolve remarks.
- Core.ScriptedAlarms-011: propagate the no-op reason through
TransitionResult.NoOp(state, reason) and log it from
ScriptedAlarmEngine.ApplyAsync.
- Core.ScriptedAlarms-009 (Won't Fix per recommendation): documented the
per-evaluation dictionary allocation in docs/v2/Galaxy.Performance.md
with a mitigation path if a future soak surfaces pressure.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Documents the four perf surfaces shipped in Phase 6:
- Tracing surface (PR 6.1) — table of every span the driver emits +
rationale for stream-level (not per-event) coverage.
- Metrics surface (PR 6.2) — three EventPump counters, tagging
scheme, the bounded-channel design, and the
received = dispatched + dropped + in-flight invariant.
- Buffered update interval (PR 6.3) — how MxAccess.PublishingIntervalMs
flows through both subscribe paths and what's still pending on the
gw side (typed SetBufferedUpdateInterval helper).
- Soak scenario (PR 6.4) — env-var-gated 24h × 50k validation with
the CI-compressed override recipe.
- Tuned defaults (PR 6.5) — table of every default with source +
notes; rows marked "unchanged" carry the explicit "no live data
argues for changing this" caveat.
Closes with a "where to look first when something's slow" runbook
section so on-call doesn't have to re-derive the trace+metric
correlation map from primary docs.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
