[codex] Tame nightly CI gates

.github/workflows/official-suite-sharded.yml

@@ -139,7 +139,7 @@ jobs:
     needs: [utf-shards]
     if: always()
     env:
-      UTF_GATE_MODE: ${{ github.event_name == 'schedule' && 'strict' || github.event.inputs.utf_gate_mode || 'warn' }}
+      UTF_GATE_MODE: ${{ github.event.inputs.utf_gate_mode || 'warn' }}
     steps:
       - name: Download shard 0 artifact
         uses: actions/download-artifact@v4

.github/workflows/r3-failure-ledger.yml

@@ -15,6 +15,10 @@ on:
         description: "Fail workflow when readiness streak threshold is not met (true|false)"
         required: false
         default: "false"
+      fail_on_failures:
+        description: "Fail workflow when the generated ledger contains failures (true|false)"
+        required: false
+        default: "false"
   schedule:
     - cron: "30 4 * * *"
 
@@ -33,6 +37,7 @@ jobs:
       OFFICIAL_SPECS_REF: ${{ github.event.inputs.official_specs_ref || '99704fa8860777da1d919ef765af1e41e75f5859' }}
       READINESS_MIN_STREAK: ${{ github.event.inputs.readiness_min_streak || '3' }}
       FAIL_ON_READINESS: ${{ github.event.inputs.fail_on_readiness || 'false' }}
+      FAIL_ON_FAILURES: ${{ github.event.inputs.fail_on_failures || 'false' }}
       MONGO_CONTAINER_NAME: jongodb-replset
       MONGO_REPLSET_NAME: rs0
       MONGO_PORT: "27017"
@@ -74,7 +79,7 @@ jobs:
             -Pr3SpecRepoRoot="${RUNNER_TEMP}/mongodb-specifications" \
             -Pr3FailureLedgerOutputDir="${R3_LEDGER_OUTPUT_DIR}" \
             -Pr3FailureLedgerMongoUri="${JONGODB_REAL_MONGOD_URI}" \
-            -Pr3FailureLedgerFailOnFailures=true \
+            -Pr3FailureLedgerFailOnFailures="${FAIL_ON_FAILURES}" \
             r3FailureLedger
 
       - name: Build trend snapshot
@@ -515,7 +520,7 @@ jobs:
               raise SystemExit(1)
           PY
 
-      - name: Post triage breadcrumb on schedule failure
+      - name: Post triage breadcrumb on scheduled ledger findings
         if: always() && github.event_name == 'schedule'
         env:
           GH_TOKEN: ${{ github.token }}

.github/workflows/runon-lane-pilot.yml

@@ -187,14 +187,15 @@ jobs:
                   handle.write(markdown)
                   handle.write("\n")
 
+          warnings = []
           failures = []
           if strict_mismatch != 0 or strict_error != 0:
-              failures.append(
-                  f"strict lane stability gate failed: mismatch={strict_mismatch}, error={strict_error}"
+              warnings.append(
+                  f"strict lane has non-zero mismatch/error: mismatch={strict_mismatch}, error={strict_error}"
               )
           if extended_mismatch > strict_mismatch or extended_error > strict_error:
-              failures.append(
-                  "extended lane regression gate failed: "
+              warnings.append(
+                  "extended lane has higher mismatch/error after importing more runOn cases: "
                   f"strict(mismatch={strict_mismatch}, error={strict_error}) vs "
                   f"extended(mismatch={extended_mismatch}, error={extended_error})"
               )
@@ -207,6 +208,9 @@ jobs:
                   f"coverage gate failed: extended imported={extended_imported} is lower than strict={strict_imported}"
               )
 
+          if warnings:
+              for message in warnings:
+                  print(f"::warning::{message}")
           if failures:
               for message in failures:
                   print(f"::error::{message}")
@@ -325,15 +329,16 @@ jobs:
                   / strict_totals["runOnNotSatisfied"]
               ) * 100.0
 
+          warnings = []
           failures = []
           if strict_totals["mismatch"] != 0 or strict_totals["error"] != 0:
-              failures.append(
-                  "strict lane aggregate stability gate failed: "
+              warnings.append(
+                  "strict lane aggregate has non-zero mismatch/error: "
                   f"mismatch={strict_totals['mismatch']}, error={strict_totals['error']}"
               )
           if extended_totals["mismatch"] > strict_totals["mismatch"] or extended_totals["error"] > strict_totals["error"]:
-              failures.append(
-                  "extended lane aggregate regression gate failed: "
+              warnings.append(
+                  "extended lane aggregate has higher mismatch/error after importing more runOn cases: "
                   f"strict(mismatch={strict_totals['mismatch']}, error={strict_totals['error']}) vs "
                   f"extended(mismatch={extended_totals['mismatch']}, error={extended_totals['error']})"
               )
@@ -361,6 +366,7 @@ jobs:
               "shards": shard_rows,
               "gatePassed": not failures,
               "gateFailures": failures,
+              "gateWarnings": warnings,
           }
           summary_json.parent.mkdir(parents=True, exist_ok=True)
           summary_json.write_text(json.dumps(summary, ensure_ascii=False, indent=2) + "\n", encoding="utf-8")
@@ -394,6 +400,11 @@ jobs:
               lines.append("## Gate failures")
               for failure in failures:
                   lines.append(f"- {failure}")
+          if warnings:
+              lines.append("")
+              lines.append("## Gate warnings")
+              for warning in warnings:
+                  lines.append(f"- {warning}")
 
           markdown = "\n".join(lines) + "\n"
           summary_md.write_text(markdown, encoding="utf-8")
@@ -405,6 +416,9 @@ jobs:
                   handle.write(markdown)
                   handle.write("\n")
 
+          if warnings:
+              for warning in warnings:
+                  print(f"::warning::{warning}")
           if failures:
               for failure in failures:
                   print(f"::error::{failure}")

docs/USAGE.md

@@ -138,8 +138,8 @@ Current certified reference runs (as of 2026-02-28):
 - R3 External Canary Certification (latest success): `22378993613` (2026-02-25)
 
 Official Suite Sharded UTF gate modes:
-- `warn` (default for `pull_request` and manual dispatch): publish aggregate `total/match/mismatch/error` + pass rate summary, emit warning when mismatch/error remain.
-- `strict` (default for scheduled `main` run): fail workflow when aggregate mismatch/error is non-zero.
+- `warn` (default for `pull_request`, scheduled `main`, and manual dispatch): publish aggregate `total/match/mismatch/error` + pass rate summary, emit warning when mismatch/error remain.
+- `strict`: fail workflow when aggregate mismatch/error is non-zero. Use this for release-candidate certification, not routine nightly signal collection.
 - `off`: summary only, no mismatch/error gate.
 - Manual dispatch input: `utf_gate_mode` (`off | warn | strict`).
 - Workflow streak artifact: `utf-shard-streak.json/.md` exposing `officialZeroMismatchStreak` over recent schedule history.
@@ -208,6 +208,7 @@ R3 Failure Ledger workflow (`.github/workflows/r3-failure-ledger.yml`) also:
 - manual dispatch inputs:
   - `readiness_min_streak` (default `3`)
   - `fail_on_readiness` (`true|false`, default `false`)
+  - `fail_on_failures` (`true|false`, default `false`; set `true` for release-candidate certification)
 
 Run Spring compatibility matrix: