[SVLS-9175] feat: emit OOM metric on memory equality with per-request dedup#1241

Merged

lym953 merged 22 commits into

mainfrom

yiming.luo/fix-1237-node-oom-metric

Jun 5, 2026

Commits on May 29, 2026

fix(metrics): emit OOM metric when max_memory_used == memory_size, with per-request dedup

lym953
and
claude
committed
test(integration): add cross-runtime OOM test suite

lym953
and
claude
committed

Commits on Jun 2, 2026

chore: review comment cleanup and Ruby layer bump

lym953
and
claude
committed
chore: add debug logs for OOM metric emit-without-dedup branches

lym953
and
claude
committed
test(integration): add LMI OOM integration test (Python)

lym953
and
claude
committed
docs: correct LMI OOM characterization based on empirical verification

lym953
and
claude
committed
docs: make best-effort nature of OOM metric dedup explicit

lym953
and
claude
committed
fix(logs): prefer log-payload requestId for OOM events in LMI mode

lym953
and
claude
committed
refactor(logs): drop current_request_id() helper, use payload requestId universally

lym953
and
claude
committed
Revert "refactor(logs): drop current_request_id() helper, use payload requestId universally"
lym953
committed
fix(logs): always extract payload requestId; keep current_request_id() fallback

lym953
and
claude
committed
fix(integration): LMI OOM stack — meet 2 GB minimum, use 100 GB single-shot allocator

lym953
and
claude
committed
fix(integration): subtract 60s from OOM query windowStart to include rollup bucket

lym953
and
claude
committed
docs: update oom-go comment — 192 MB → 256 MB

lym953
and
claude
committed

Commits on Jun 4, 2026