weichen/Agent-System
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(ci): stabilize FullNone PI E2E timeouts and host KB DNS

Browse Source 
Raise HTTP client timeout above AgentTimeout, rewrite compose KB host
to localhost for host-side tests, and retry nightly on deadline flakes.

Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
weichen
2026-09-17 00:28:38 +00:00
co-authored by Cursor
parent 4f2e5f013f
commit a02c67db82
4 changed files with 12 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
internal/server/piagent_e2e_test.go
+5
View File
@@ -40,6 +40,11 @@ func mergePiAgentE2EKbConfig(cfg *config.Config) {
if cfg.KbDbHost == "" {
cfg.KbDbHost = "127.0.0.1"
}
// Host-side go test cannot use compose DNS; shuimu-kb-shared often resolves to a
// poisoned/VPN address (e.g. 198.18.0.0/15). Prefer the published localhost port.
if cfg.KbDbHost == "shuimu-kb-shared" || cfg.KbDbHost == "shuimu-kb" {
cfg.KbDbHost = "127.0.0.1"
}
if cfg.KbDbPort == 0 {
cfg.KbDbPort = 5432
}
internal/server/workflow_governance_full_none_piagent_pg_test.go
+3 -1
View File
@@ -39,7 +39,9 @@ func TestWorkflowGovernance_PiAgent_PG_FullNoneMixedChain(t *testing.T) {
handler, srv := setupPGPiAgentServer(t, pool)
ts := httptest.NewServer(handler)
defer ts.Close()
client := &http.Client{Timeout: 50 * time.Minute}
// Must exceed setupPGPiAgentServer AgentTimeout (60m) so the client does not
// cancel while the server is still waiting on the PI runner.
client := &http.Client{Timeout: 75 * time.Minute}
seedOfficialWorkflowManifest(t, pool, workflowdef.OfficialKeyMain)
internal/server/workflow_governance_piagent_pg_test.go
+1 -1
View File
@@ -79,7 +79,7 @@ func runPiAgentGovernanceE2E(t *testing.T, spec piAgentGovernanceCase) {
handler, srv := setupPGPiAgentServer(t, pool)
ts := httptest.NewServer(handler)
defer ts.Close()
client := &http.Client{Timeout: 50 * time.Minute}
client := &http.Client{Timeout: 75 * time.Minute}
seedOfficialWorkflowManifest(t, pool, spec.WfKey)
scripts/run-nightly-real-pi.sh
+3 -3
View File
@@ -135,9 +135,9 @@ LOG_FILE="${EVIDENCE_DIR}/go-test.txt"
set +e
go test -count=1 -timeout "${TIMEOUT}" -v ./internal/server/ -run "${RUN_FILTER}" 2>&1 | tee "${LOG_FILE}"
TEST_RC=${PIPESTATUS[0]}
# One retry when model quality_failed flakes the U8 generate (product path, not infra).
if [[ "${TEST_RC}" -ne 0 ]] && grep -Eq 'quality_failed want completed|status=quality_failed' "${LOG_FILE}"; then
echo "==> retry once after quality_failed flake..." | tee -a "${LOG_FILE}"
# One retry for known flakes: model quality_failed, or PI hang / client deadline.
if [[ "${TEST_RC}" -ne 0 ]] && grep -Eq 'quality_failed want completed|status=quality_failed|context deadline exceeded|Client\.Timeout exceeded' "${LOG_FILE}"; then
echo "==> retry once after quality_failed/deadline flake..." | tee -a "${LOG_FILE}"
RETRY_LOG="${EVIDENCE_DIR}/go-test-retry.txt"
go test -count=1 -timeout "${TIMEOUT}" -v ./internal/server/ -run "${RUN_FILTER}" 2>&1 | tee "${RETRY_LOG}"
TEST_RC=${PIPESTATUS[0]}