test: prove pg-erd runtime isolation under real traffic - #16
Conversation
|
Important Draft PR not reviewedDraft PRs are not automatically reviewed by default.
To automatically review draft PRs, update your CodeRabbit configuration: reviews:
auto_review:
drafts: trueThanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
seonghobae
added
documentation
ChatGPT Codex Connector
|
@coderabbitai review Please review exact current parent |
|
I will treat prior
|
|
@coderabbitai review Fresh exact-head review requested for |
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
base="f3dbe7afe1fa945d74145bb7852ae77882145b0a"
head="d16f1440a05ca92a2acacd247ae2e62644df4f22"
echo "== Exact commit topology =="
git show -s --format='commit=%H%nparents=%P%nsubject=%s' "$head"
git merge-base "$base" "$head"
git rev-list --left-right --count "$base...$head"
git log --format='%H %P %s' --ancestry-path "$base..$head"
echo
echo "== Exact effective paths and summary =="
git diff --name-status "$base" "$head"
git diff --stat "$base" "$head"
echo
echo "== Changed-file structure =="
for path in $(git diff --name-only "$base" "$head"); do
echo "--- $path ---"
case "$path" in
*.rs) ast-grep outline "$path" --items all --view expanded || true ;;
*) sed -n '1,260p' "$path" ;;
esac
done
echo
echo "== Relevant production and test symbols =="
rg -n -C 3 --glob '*.rs' \
'MigrationGatewayProxy|cwl-pingora-pg-erd-migration|max_in_flight_requests|backpressure_rejections_total|readyz|body.*budget|Body.*Limit|413|503' \
. || trueLength of output: 87559 🏁 Script executed: #!/bin/bash
set -euo pipefail
base="f3dbe7afe1fa945d74145bb7852ae77882145b0a"
merge="e087639ad2e03d36acdcb634d5081bd0b6cab343"
head="d16f1440a05ca92a2acacd247ae2e62644df4f22"
echo "== Two-parent adoption verification =="
git show -s --format='merge=%H%nparents=%P%ntree=%T%nsubject=%s' "$merge"
echo "base tree: $(git rev-parse "$base^{tree}")"
echo "merge tree: $(git rev-parse "$merge^{tree}")"
echo "Differences: exact parent -> adoption merge tree"
git diff --name-status "$base" "$merge"
echo "Differences: adoption merge tree -> exact head"
git diff --name-status "$merge" "$head"
echo
echo "== Full new traffic acceptance =="
cat -n tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Migration callback lifecycle =="
sed -n '105,270p' src/migration_proxy.rs
echo
echo "== Existing compiled pg-erd fixture synchronization reference =="
sed -n '1,430p' tests/pg_erd_production_path.rs
echo
echo "== Exact diff for the added test =="
git diff --no-ext-diff --unified=80 "$base" "$head" -- tests/pg_erd_runtime_isolation_traffic.rsLength of output: 43078
The effective diff has only the five stated paths. The adoption merge changes only the dedicated traffic test and three code-current documents. The head commit then changes only I found one acceptance-test gap:
The test keeps product authentication, business behavior, Keyverse identity, and Wardnet/EgressWeave policy outside the gateway boundary. CI You are interacting with an AI system. |
|
@coderabbitai review Please review exact current head |
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
base="bb65f2810b178bda46ed7eb5c6a62aae9fe36403"
head="4e6376404986cd9c94a0894c5d5e5ba51aeb9f9c"
red="d16f1440a05ca92a2acacd247ae2e62644df4f22"
repair="abee6af1ecfc1d2f44de1dc25917c2eb84a6bb38"
echo "== Exact topology =="
git show -s --format='head=%H%nparents=%P%ntree=%T%nsubject=%s' "$head"
echo "base_tree=$(git rev-parse "$base^{tree}")"
echo "merge_base=$(git merge-base "$base" "$head")"
echo "range_count=$(git rev-list --left-right --count "$base...$head")"
git log --format='%H %P %s' --ancestry-path "$base..$head"
echo
echo "== Effective paths: exact parent -> exact head =="
git diff --name-status "$base" "$head"
git diff --stat "$base" "$head"
echo
echo "== Repair-only comparison =="
git diff --no-ext-diff --stat "$red" "$repair"
git diff --no-ext-diff --word-diff=porcelain "$red" "$repair" -- tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Changed Rust test outline and contents =="
ast-grep outline tests/pg_erd_runtime_isolation_traffic.rs --items all --view expanded || true
cat -n tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Relevant production/test references =="
rg -n -C 4 --glob '*.rs' \
'cwl-pingora-pg-erd-migration|MigrationGatewayProxy|max_in_flight_requests|backpressure_rejections_total|readyz|body.*limit|body.*budget|413|503|retry' \
src tests || true
echo
echo "== Relevant migration proxy implementation =="
if [ -f src/migration_proxy.rs ]; then
ast-grep outline src/migration_proxy.rs --items all --view expanded || true
sed -n '1,360p' src/migration_proxy.rs
fi
echo
echo "== Exact test diff =="
git diff --no-ext-diff --unified=100 "$base" "$head" -- tests/pg_erd_runtime_isolation_traffic.rsLength of output: 104552
The head is an ordinary two-parent merge. Its first parent is The effective diff contains only these five paths:
The repair from I found one issue:
The test starts You are interacting with an AI system. |
|
@coderabbitai review 현재 exact head |
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
base="bb65f2810b178bda46ed7eb5c6a62aae9fe36403"
head="4e6376404986cd9c94a0894c5d5e5ba51aeb9f9c"
echo "== Exact topology and range =="
git show -s --format='head=%H%nparents=%P%ntree=%T%nsubject=%s' "$head"
printf 'base_tree='
git rev-parse "$base^{tree}"
printf 'merge_base='
git merge-base "$base" "$head"
printf 'left_right_count='
git rev-list --left-right --count "$base...$head"
git log --format='%H %P %s' --ancestry-path "$base..$head"
echo
echo "== Effective paths and patch summary =="
git diff --name-status "$base" "$head"
git diff --stat "$base" "$head"
git diff --check "$base" "$head"
echo
echo "== Traffic acceptance test =="
ast-grep outline tests/pg_erd_runtime_isolation_traffic.rs --items all --view expanded || true
cat -n tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Exact test delta =="
git diff --no-ext-diff --unified=120 "$base" "$head" -- tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Relevant Rust symbols and call contexts =="
rg -n -C 5 --glob '*.rs' \
'cwl-pingora-pg-erd-migration|MigrationGatewayProxy|max_in_flight_requests|backpressure_rejections_total|readyz|Body.*[Ll]imit|body.*[Bb]udget|413|503|timeout|Timeout|retry|Retry' \
src tests || true
echo
echo "== Candidate implementation outlines =="
for f in $(fd -t f -e rs . src tests | sort); do
if rg -q 'MigrationGatewayProxy|backpressure_rejections_total|max_in_flight_requests|readyz' "$f"; then
echo "--- $f ---"
ast-grep outline "$f" --items all --view expanded || true
fi
doneLength of output: 164780 🏁 Script executed: #!/bin/bash
set -euo pipefail
echo "== Migration proxy callbacks =="
sed -n '105,270p' src/migration_proxy.rs
echo
echo "== Fixed pg-erd routing plan and delivery lookup =="
ast-grep outline src/migration_plan.rs src/migration_delivery.rs --items all --view expanded || true
sed -n '1,280p' src/migration_plan.rs
sed -n '1,260p' src/migration_delivery.rs
echo
echo "== pg-erd production fixture lifecycle reference =="
sed -n '1,430p' tests/pg_erd_production_path.rs
echo
echo "== Exact inherited-parent drift checks =="
base="bb65f2810b178bda46ed7eb5c6a62aae9fe36403"
head="4e6376404986cd9c94a0894c5d5e5ba51aeb9f9c"
git diff --name-only "$base" "$head" | while IFS= read -r path; do
printf '%s: ' "$path"
git diff --quiet "$base" "$head" -- "$path" || echo changed
done
git diff --name-status "$base" "$head" -- src .github Dockerfile Cargo.toml Cargo.lock 2>/dev/null || trueLength of output: 34762
범위는 ahead 16 / behind 0입니다. head는 일반적인 두 부모 merge입니다. 유효 변경 경로는 다음 다섯 개뿐입니다.
상속된 한 가지 문제를 확인했습니다.
fixture는 시작 대기, channel 대기, downstream read에 timeout을 사용합니다. 다만 recovery fixture의 You are interacting with an AI system. |
|
@coderabbitai review Fresh exact-head review for |
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
base="bb65f2810b178bda46ed7eb5c6a62aae9fe36403"
head="356f3f250043d71c9bb1c0655481cd6984f3ae57"
echo "== Exact topology =="
git show -s --format='head=%H%nparents=%P%ntree=%T%nsubject=%s' "$head"
printf 'base_tree='; git rev-parse "$base^{tree}"
printf 'merge_base='; git merge-base "$base" "$head"
printf 'ahead_behind='; git rev-list --left-right --count "$base...$head"
git log --format='%H %P %s' --ancestry-path "$base..$head"
echo
echo "== Effective paths and diff hygiene =="
git diff --name-status "$base" "$head"
git diff --stat "$base" "$head"
git diff --check "$base" "$head"
echo
echo "== Changed Rust test outline and complete source =="
ast-grep outline tests/pg_erd_runtime_isolation_traffic.rs --items all --view expanded
cat -n tests/pg_erd_runtime_isolation_traffic.rs
echo
echo "== Exact test and docs deltas =="
git diff --no-ext-diff --unified=120 "$base" "$head" -- \
tests/pg_erd_runtime_isolation_traffic.rs \
docs/product-technical-gap-baseline.md \
CHANGELOG.md CLAUDE.md TEST_STRATEGY.md
echo
echo "== Relevant implementation map =="
for file in src/migration_proxy.rs src/migration_plan.rs src/migration_delivery.rs; do
if [ -f "$file" ]; then
echo "--- $file ---"
ast-grep outline "$file" --items all --view expanded
fi
done
echo
echo "== Relevant implementation and fixture contexts =="
rg -n -C 8 --glob '*.rs' \
'cwl-pingora-pg-erd-migration|MigrationGatewayProxy|max_in_flight_requests|backpressure_rejections_total|readyz|read_ms|body.*[Bb]udget|[Bb]ody.*[Ll]imit|413|503|Instant::now|timeout\(|accept\(|retry' \
src tests
echo
echo "== Exact metric assertion forms =="
rg -n -C 4 --glob '*.rs' \
'cwl_pingora_gateway_backpressure_rejections_total|metrics_body|metrics.*contains|lines\(\)' \
tests srcLength of output: 231936 🏁 Script executed: #!/bin/bash
set -euo pipefail
echo "== Migration callback ordering and lease lifetime =="
sed -n '169,268p' src/migration_proxy.rs
echo
echo "== Request admission release implementation =="
sed -n '74,116p' src/runtime_isolation.rs
echo
echo "== Compiled pg-erd entry point and metrics listener setup =="
fd -t f -e rs . src | sort | while IFS= read -r file; do
if rg -q 'cwl-pingora-pg-erd-migration|metrics_listener|build_proxy\(\)' "$file"; then
echo "--- $file ---"
rg -n -C 6 'cwl-pingora-pg-erd-migration|metrics_listener|build_proxy\(\)' "$file"
fi
doneLength of output: 13289
I found no actionable issue. The merge base is the requested
The prior fail-fast finding is closed. The metrics assertion is non-vacuous. It splits the metrics response into complete lines and compares each line for exact equality with: Therefore, samples with values such as The compiled
The fixture has bounded process startup, downstream read, and synchronization waits. The backend header read and recovery The changed test and documentation do not add product authentication, product business logic, Keyverse identity authority, Wardnet/EgressWeave authority, arbitrary route configuration, or retry authority. The exact head still requires its own hosted execution evidence. You are interacting with an AI system. |
|
@coderabbitai review Fresh exact-head review requested for current parent |
|
Tip For best results, initiate chat on the files or code changes. 🧩 Analysis chain🏁 Script executed: #!/bin/bash
set -euo pipefail
base="bb65f2810b178bda46ed7eb5c6a62aae9fe36403"
head="356f3f250043d71c9bb1c0655481cd6984f3ae57"
echo "== Topology and range =="
git show -s --format='head=%H%nparents=%P%ntree=%T%nsubject=%s' "$head"
printf 'merge_base='; git merge-base "$base" "$head"
printf 'ahead_behind='; git rev-list --left-right --count "$base...$head"
git log --format='%H %P %s' --ancestry-path "$base..$head"
echo "== Exact effective scope =="
git diff --name-status "$base" "$head"
git diff --stat "$base" "$head"
git diff --check "$base" "$head"
echo "== Protected inherited paths changed in exact range (should be empty) =="
git diff --name-status "$base" "$head" -- src .github Dockerfile Cargo.toml Cargo.lock || true
echo "== Changed test structure and contents =="
ast-grep outline tests/pg_erd_runtime_isolation_traffic.rs --items all --view expanded
cat -n tests/pg_erd_runtime_isolation_traffic.rs
echo "== Exact test patch =="
git diff --no-ext-diff --unified=100 "$base" "$head" -- tests/pg_erd_runtime_isolation_traffic.rs
echo "== Relevant implementation maps =="
for file in src/migration_proxy.rs src/runtime_isolation.rs src/migration_plan.rs src/migration_delivery.rs; do
if [ -f "$file" ]; then
echo "--- $file ---"
ast-grep outline "$file" --items all --view expanded
fi
done
echo "== Admission, health, routing, and timeout contexts =="
rg -n -C 10 --glob '*.rs' \
'cwl-pingora-pg-erd-migration|MigrationGatewayProxy|max_in_flight_requests|backpressure_rejections_total|/readyz|read_ms|body.*[Bb]udget|[Bb]ody.*[Ll]imit|Instant::now|timeout\(|accept\(|RequestAdmission|retry' \
src tests
echo "== Relevant implementations =="
sed -n '150,290p' src/migration_proxy.rs
sed -n '1,160p' src/runtime_isolation.rs
echo "== Named GitHub Actions runs =="
gh run view 34168471195 --json databaseId,status,conclusion,headSha,name,url 2>&1 || true
gh run view 34168471176 --json databaseId,status,conclusion,headSha,name,url 2>&1 || trueLength of output: 218151
I found no actionable issue. The merge base is the requested parent. The head is 19 commits ahead and 0 commits behind. The ancestry retains the ordinary two-parent adoption at The effective diff changes only these five paths:
The exact range does not change
The streamed-body test sends two chunked body fragments of 4 and 5 bytes. It requires HTTP 413 for the 9-byte body. It then requires HTTP 200 from The saturation test holds the only admitted routed request at the backend. It measures the overloaded request from immediately before The telemetry assertion splits the metrics response into lines. It compares each line exactly with: Values such as The test requires
The changed test adds no product authentication, business logic, Keyverse identity authority, Wardnet/EgressWeave authority, arbitrary route authority, or retry authority. CI run You are interacting with an AI system. |
1 participant
Dedicated runtime-isolation traffic evidence
This Draft owns only the
cwl-pingora-pg-erd-migrationreal-listener runtime-isolation acceptance that does not transfer from the generic gateway. It adds no product authentication/business logic, retry/failover policy, Keyverse identity, or Wardnet/EgressWeave authority.Exact base is #15
bb65f2810b178bda46ed7eb5c6a62aae9fe36403; exact head is356f3f250043d71c9bb1c0655481cd6984f3ae57. The effective child scope remains five paths:CHANGELOG.md,CLAUDE.md,TEST_STRATEGY.md,docs/product-technical-gap-baseline.md, andtests/pg_erd_runtime_isolation_traffic.rs.The compiled migration process is exercised over real loopback listeners. A 9-byte chunked body against an 8-byte budget must return HTTP 413 while
/readyzstays 200. Withmax_in_flight_requests=1, one routed backend request holds the sole admission lease; the next routed request must return HTTP 503 in<1 s, materially before the configured 2 s upstream read budget;/readyzstays 200; metrics must contain the exact Prometheus samplecwl_pingora_gateway_backpressure_rejections_total 1; releasing the held request must restore routed capacity.The
<1 sassertion is the causal repair for the valid predecessor finding that only eventual 503 had been proved. Exact-line metric matching is the causal repair for the prefix false-pass class (10/11matching1). Neither repair weakens traffic volume, timeout budgets or the gateway admission policy.Exact-head evidence
CI
34168471195is terminal GREEN on exact356f3f25...: load, formatting/locked compile/test/strict Clippy/warning-denied rustdoc/100% owned-production coverage/dependency-lock, and dual-profile OCI runtime all passed. The exact generic k6 artifact records 400 requests, 800/800 checks, zero HTTP failures and p95 about0.919 msagainst the unchanged<20 msloopback threshold. This remains controlled generic loopback evidence, not routed pg-erd or production SLO proof.Supply Chain
34168471176is terminal GREEN on the same exact SHA: committed dependency audit, both admitted image builds, SPDX dependency SBOM, generic and pg-erd vulnerability scans, exact-source binding and success evidence upload all completed successfully.Fresh exact-range CodeRabbit review of
bb65f281...356f3f25reports no actionable issue and confirms the fail-fast and exact-metric findings are closed; that is technical review, not independent human approval.Keep Draft. Supplier promotion remains independently blocked by #54's intentional
derivative 2.2.0RED and #56's independent approval gate. No protected merge, immutable release, shadow/canary, cutover, rollback or legacy-removal credit is claimed. Child #17 may ordinary/non-force adopt this exact GREEN parent, but #16 evidence does not transfer to the changed child head.