store: make the buffer pool sense its own state and correct from it
El SDK CI - dev / build-and-test (pull_request) Failing after 14m35s

Follow-on to the edge write barrier. That fix removed the full-store walk;
this one makes the pool able to notice if anything like it happens again.

WHAT WENT WRONG, precisely: the pool thrashed the live engram to a standstill
twice on 2026-08-15 and said nothing. From outside it was indistinguishable
from "busy loading" — 100% CPU, flat RSS, no output — so four wrong theories
got tried (bad binary, corrupt snapshot, WAL replay, feature flags), each
costing a deploy or a rollback. The whole time, hits/misses/evictions were
already being counted in PgCache, and the struct comment read:

    /* stats (introspection only — never affect semantics) */

That comment was the bug. Self-measurement treated as decoration is why the
pool could not correct itself and why no one outside could see what it was
doing. A system that cannot read its own state cannot correct, and neither can
anyone watching it.

  - pc_adapt_budget(): the loop, closed. Over a sliding window, evictions
    running at a large fraction of accesses WHILE reuse is real means the
    working set exceeds the budget — so grow it, geometrically, bounded by a
    LIVE re-read of physical memory. Evictions alone are not pressure (a scan
    evicts and never returns); evictions with reuse are. An explicit
    ENGRAM_POOL_FRAMES still wins — an operator override must not be silently
    overruled.

  - Budget derived, not declared. A constant cannot be right: 16 GiB of frames
    is arbitrary on a 48 GB host and suicidal on a 16 GB one. Even "60% of RAM
    at startup" is a guess about the future — it cannot know the store grew or
    the machine changed. Hence the live re-read.

  - pc_report(): ONE structured emission carrying the entire sensed state,
    through emit_log — El's existing telemetry, already exporting to OTLP.
    Deliberately not a function per stat, and deliberately not a bespoke
    /api/pool endpoint: both make observability something hand-written per noun
    instead of the uniform mechanism every component already has.

  - engram_pool_stats_json(): the same state readable live, wired through the
    normal builtin path (codegen arity + el_seed wrapper), so the pool can be
    observed in real time rather than reconstructed afterward from a stack
    sample.

Verified: with the exact configuration that took production down
(ENGRAM_POOL_FRAMES=65536 → 1 GiB cache against a 2 GiB store) the engram boots
clean and serves — 0.0% CPU, 13,436 nodes / 37,663 edges, embeddings complete —
and NO pressure event fires, because the barrier removed the walk that caused
it. The controller is defense in depth; the barrier is the fix.
This commit is contained in:
bigmerge
2026-08-15 20:44:23 -05:00
parent 777ccc02f0
commit e917b3d439
5 changed files with 157 additions and 27 deletions
+2
View File
@@ -2765,6 +2765,7 @@ fn builtin_arity(name: String) -> Int {
if str_eq(name, "__engram_connect_in") { return 5 }
if str_eq(name, "__engram_scan_nodes_json") { return 2 }
if str_eq(name, "__engram_edges_json") { return 2 }
if str_eq(name, "__engram_pool_stats_json") { return 0 }
if str_eq(name, "__generate") { return 1 }
// Filesystem
if str_eq(name, "fs_read") { return 1 }
@@ -2864,6 +2865,7 @@ fn builtin_arity(name: String) -> Int {
if str_eq(name, "engram_search_json") { return 2 }
if str_eq(name, "engram_scan_nodes_json") { return 2 }
if str_eq(name, "engram_edges_json") { return 2 }
if str_eq(name, "engram_pool_stats_json") { return 0 }
if str_eq(name, "engram_neighbors_json") { return 3 }
if str_eq(name, "engram_activate_json") { return 2 }
if str_eq(name, "engram_stats_json") { return 0 }
+38
View File
@@ -18373,3 +18373,41 @@ el_val_t engram_edges_json(el_val_t limit, el_val_t offset) {
jb_putc(&b, ']');
return el_wrap_str(b.buf);
}
/* engram_pool_stats_json() — the buffer pool's interoception, exposed.
*
* StorePoolStats and store_pool_stats() already existed and were surfaced
* NOWHERE. On 2026-08-15 the engram thrashed itself to a standstill twice while
* these exact counters sat in memory, unread, and four wrong theories were tried
* from the outside instead. Sensing state is only corrective if the state can be
* read by the process itself (pc_adapt_budget) and by anything watching it.
*
* Serves the live numbers plus the derived signals that actually diagnose:
* hit_rate sustained low hit rate with high evictions is the thrash shape
* evict_ratio evictions per access; ~1 means every fetch displaces a live page
* pressure 1 when evicting into genuine reuse (working set > budget)
* cap_gib/resident_gib budget vs what is actually held
*/
el_val_t engram_pool_stats_json(void) {
if (!g_engram_store) return el_wrap_str(el_strdup("{\"store\":false}"));
StorePoolStats st;
store_pool_stats(g_engram_store, &st);
uint64_t acc = st.hits + st.misses;
double hit_rate = acc ? (double)st.hits / (double)acc : 0.0;
double evict_ratio = acc ? (double)st.evictions / (double)acc : 0.0;
int pressure = (acc > 100000 && evict_ratio > 0.33 && hit_rate > 0.25) ? 1 : 0;
char b[768];
snprintf(b, sizeof b,
"{\"store\":true,\"cap_frames\":%zu,\"resident_frames\":%zu,\"pinned\":%zu,"
"\"dirty\":%zu,\"prefetch\":%u,\"hits\":%llu,\"misses\":%llu,\"evictions\":%llu,"
"\"prefetch_reads\":%llu,\"hit_rate\":%.4f,\"evict_ratio\":%.4f,\"pressure\":%d,"
"\"cap_gib\":%.3f,\"resident_gib\":%.3f,\"page_size\":%u}",
st.cap, st.resident, st.pinned, st.dirty, st.prefetch,
(unsigned long long)st.hits, (unsigned long long)st.misses,
(unsigned long long)st.evictions, (unsigned long long)st.prefetch_reads,
hit_rate, evict_ratio, pressure,
(double)st.cap * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0),
(double)st.resident * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0),
(unsigned)STORE_PAGE_SIZE);
return el_wrap_str(el_strdup(b));
}
+3
View File
@@ -1025,6 +1025,9 @@ el_val_t engram_recall_json(el_val_t query, el_val_t limit);
* whole-graph round trip that /api/graph/edges used to do. */
el_val_t engram_edges_json(el_val_t limit, el_val_t offset);
/* Buffer-pool interoception as JSON — live pool health for observation. */
el_val_t engram_pool_stats_json(void);
/* CGI identity accessors (read-only). */
el_val_t cgi_principal(void);
el_val_t cgi_network(void);
+3
View File
@@ -1376,6 +1376,9 @@ el_val_t __engram_edges_json(el_val_t limit, el_val_t offset) {
return engram_edges_json(limit, offset);
}
el_val_t engram_pool_stats_json(void);
el_val_t __engram_pool_stats_json(void) { return engram_pool_stats_json(); }
el_val_t __engram_scan_nodes_by_type_json(el_val_t node_type, el_val_t limit, el_val_t offset) {
return engram_scan_nodes_by_type_json(node_type, limit, offset);
}
+111 -27
View File
@@ -239,8 +239,15 @@ struct PgCache {
unsigned prefetch; /* read-ahead window (pages); 0 = off */
LayerPin* lp; size_t lp_n, lp_cap; /* hot-layer pin bookkeeping */
size_t dirty_count; /* # dirty frames, maintained incrementally (M5) */
/* stats (introspection only — never affect semantics) */
/* Interoception. These were "introspection only — never affect semantics",
* and that was the bug: the pool could not feel itself thrash, so it could
* not correct, and neither could anyone watching from outside. The sensed
* state IS the corrective mechanism (see pc_adapt_budget) the same way the
* engram's own boundary-beat/chronoception let it feel its own activity. */
uint64_t hits, misses, evictions, prefetch_reads;
/* sliding-window marks so pressure reflects NOW, not lifetime totals */
uint64_t adapt_last_acc, adapt_last_evic, adapt_last_hits;
uint64_t adapt_grows; /* how many times the budget corrected upward */
};
/* ── little-endian scalar codecs ──────────────────────────────────────────── */
@@ -1781,35 +1788,112 @@ static void pc_remove(PgCache* c, PgEnt* e){
/* Reclaim clean unpinned frames from the LRU end until under budget, or until no
* evictable frame remains (a dirty/pinned-heavy pool may transiently exceed cap
* that is the no-steal guarantee, not a bug: the next checkpoint frees them). */
/* Thrash detector.
/* ── Adaptive budget: close the loop ─────────────────────────────────────────
*
* Thrashing is not slowness; it is zero progress, and from the outside it is
* indistinguishable from "loading a big store" 100% CPU, flat RSS, no output.
* That ambiguity cost hours on 2026-08-15: the process was assumed to be busy
* when it was in fact evicting each page moments before needing it again.
* THE LESSON THIS ENCODES (2026-08-15). The engram spent hours down while four
* separate theories were tried bad binary, corrupt snapshot, WAL replay,
* feature flags because nothing in the system said what was happening. It
* looked identical to "busy loading": 100% CPU, flat RSS, no output. Meanwhile
* hits/misses/evictions were ALREADY being counted, right here, and surfaced
* nowhere. One eviction-rate number would have ended it in seconds.
*
* The signature is unambiguous and cheap to watch: evictions climbing at a rate
* comparable to accesses, i.e. nearly every fetch pushing out a live frame.
* A cache doing useful work evicts far less often than it hits. Say so, once,
* loudly, with the numbers and the remedy silence here is what made this
* expensive to find. */
static void pc_thrash_check(PgCache* c){
static int warned = 0;
if (warned) return;
* So the counters are not decoration. They are the control signal.
*
* A budget chosen once a literal like 65536, or 60% of RAM read at startup
* is a guess about the future. It cannot know the store grew, the working set
* shifted, or another process took the memory. The cache already MEASURES the
* only thing that matters (am I evicting pages I am about to want again), so it
* should act on that measurement instead of on a number someone typed.
*
* The controller: over a sliding window, if evictions are running at a rate
* comparable to accesses AND there is genuine reuse (hits are material), the
* working set exceeds the budget grow it. Growth is geometric, bounded by a
* live re-read of physical memory rather than a value cached at boot, so it
* tracks the machine instead of a snapshot of it. It never shrinks on its own:
* cap is a ceiling, not an allocation, and frames are only ever held because a
* real access put them there.
*
* Two things this deliberately does NOT do: it does not attempt a cleverer
* eviction policy (when the working set does not fit, no policy helps that is
* Denning, and it is why "tune the LRU" was never the fix), and it does not stay
* silent (pool_report exposes the same numbers outward, so a human or a metric
* pipeline sees the pressure the controller is reacting to). */
/* El's native telemetry, already in the runtime and already exporting to OTLP.
* Declared weak so engram_store.c still links standalone; when the runtime is
* present (every real build) the pool's interoception flows into the SAME
* pipeline as every other metric.
*
* ONE emission carrying the whole sensed state not a function per stat, and
* not a bespoke per-subsystem endpoint. Both of those are the degenerate case:
* they make observability something you hand-write per noun instead of a
* uniform mechanism every component already has. el_val_t is int64_t; strings
* ride as pointers cast through it (see el_runtime.h's value model). */
__attribute__((weak)) int64_t emit_log(int64_t level, int64_t msg, int64_t fields_json);
static void pc_report(const PgCache* c, const char* cause){
if (!emit_log) return; /* runtime not linked: no-op */
uint64_t acc = c->hits + c->misses;
if (acc < 200000) return; /* need a real sample */
if (c->evictions * 2 < acc) return; /* evicting < half of accesses: healthy */
if (c->hits > c->evictions) return; /* still getting real reuse */
warned = 1;
fprintf(stderr,
"[engram] THRASHING: %llu evictions across %llu accesses (hits %llu, misses %llu) "
"with a %zu-frame budget (%.1f GiB). The working set exceeds the cache, so pages are "
"evicted just before they are reused and the store makes no forward progress. "
"Raise the budget (unset ENGRAM_POOL_FRAMES to derive it from host RAM, or raise "
"ENGRAM_POOL_MEM_PCT); a different eviction policy cannot fix this.\n",
(unsigned long long)c->evictions, (unsigned long long)acc,
char f[512];
snprintf(f, sizeof f,
"{\"component\":\"engram.pool\",\"cause\":\"%s\",\"hits\":%llu,\"misses\":%llu,"
"\"evictions\":%llu,\"prefetch_reads\":%llu,\"cap_frames\":%zu,\"resident\":%zu,"
"\"dirty\":%zu,\"grows\":%llu,\"hit_rate\":%.4f,\"evict_ratio\":%.4f,"
"\"cap_gib\":%.3f,\"resident_gib\":%.3f}",
cause,
(unsigned long long)c->hits, (unsigned long long)c->misses,
c->cap, (double)c->cap * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0));
(unsigned long long)c->evictions, (unsigned long long)c->prefetch_reads,
c->cap, c->count, c->dirty_count, (unsigned long long)c->adapt_grows,
acc ? (double)c->hits / (double)acc : 0.0,
acc ? (double)c->evictions / (double)acc : 0.0,
(double)c->cap * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0),
(double)c->count * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0));
emit_log((int64_t)(uintptr_t)"warn", (int64_t)(uintptr_t)"engram.pool pressure",
(int64_t)(uintptr_t)f);
}
static uint64_t pc_ram_bytes_live(void){ return pc_physical_ram(); }
static void pc_adapt_budget(PgCache* c){
if (!c->cap) return; /* unlimited: nothing to adapt */
if (getenv("ENGRAM_POOL_FRAMES")) return; /* explicit operator override wins */
/* Sliding window so the signal reflects NOW, not lifetime totals. */
uint64_t acc = c->hits + c->misses;
if (acc - c->adapt_last_acc < 100000) return;
uint64_t d_acc = acc - c->adapt_last_acc;
uint64_t d_evic = c->evictions - c->adapt_last_evic;
uint64_t d_hits = c->hits - c->adapt_last_hits;
c->adapt_last_acc = acc; c->adapt_last_evic = c->evictions; c->adapt_last_hits = c->hits;
/* Pressure = evicting on a large fraction of accesses while still getting
* real reuse. Evictions alone are normal (a scan evicts and never returns);
* evictions WITH reuse means the working set genuinely does not fit. */
if (d_evic * 3 < d_acc) return; /* < 1/3 of accesses evict: healthy */
if (d_hits * 4 < d_acc) return; /* little reuse: a scan, not pressure */
uint64_t ram = pc_ram_bytes_live(); /* live, not a boot-time constant */
if (!ram) return;
unsigned pct = 80; /* hard ceiling for autonomous growth */
const char* mp = getenv("ENGRAM_POOL_MAX_PCT");
if (mp && *mp){ unsigned long v = strtoul(mp, NULL, 10); if (v > 0 && v <= 95) pct = (unsigned)v; }
size_t ceiling = (size_t)(((ram / 100u) * pct) / (uint64_t)STORE_PAGE_SIZE);
if (c->cap >= ceiling) return; /* already at the machine's limit */
size_t want = c->cap + (c->cap / 2) + 1; /* ×1.5, geometric */
if (want > ceiling) want = ceiling;
size_t was = c->cap;
c->cap = want;
c->adapt_grows++;
/* Emit the sensed state, not just the reaction. These are the numbers that
* would have diagnosed 2026-08-15 in seconds instead of hours. */
pc_report(c, "budget-grow");
fprintf(stderr,
"[engram] pool pressure: %llu evictions / %llu accesses (%llu hits) at %zu frames "
"(%.2f GiB) — working set exceeds budget; growing to %zu frames (%.2f GiB).\n",
(unsigned long long)d_evic, (unsigned long long)d_acc, (unsigned long long)d_hits,
was, (double)was * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0),
c->cap,(double)c->cap * (double)STORE_PAGE_SIZE / (1024.0*1024.0*1024.0));
fflush(stderr);
}
@@ -1824,7 +1908,7 @@ static void pc_evict_to_budget(PgCache* c){
}
if (!freed) break; /* nothing evictable — allowed to exceed cap */
}
pc_thrash_check(c);
pc_adapt_budget(c);
}
static PgEnt* pc_get(EngramPagedStore* s, uint64_t id){