Evaluation Run Detail Report
Run ID: evalrun-3e8725979333434fa751c2dfb9a6b21f
Suite: core-daily-demo @ e3571509e8ac

Case Details
- component-restart-during-incident-window: verdict=fail failure=critical_evidence_missing subtype=- score=20 duration=2m41s workspace=ws-eval-1778483930424-ae89fba7
  token usage: input=866 output=1117 cache_read=43904 cache_write=0 cached_total=43904 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: fail score=0/30
    evidence routing: pass score=5/5
    diagnosis: fail score=0/50
  critical evidence: fail score=0/30
    - restart_signal_present earned=0/8 desc=Direct restart evidence is present within the incident window.
    - panic_signal_present earned=0/9 desc=Direct panic evidence is present within the incident window.
    - query_interruption_signal_present earned=0/8 desc=Direct query interruption evidence is present within the incident window.
  diagnosis: fail score=0/50
    - identify_known_bug_restart_chain earned=0/3 desc=Identifies the known-bug single-instance restart chain rather than a generic restart explanation.
    - identify_kill_query_cleanup_chain earned=0/3 desc=Identifies KILL QUERY / query interruption as trigger and explains the cleanup chain toward panic or restart.
    - mention_github_issue_id earned=0/3 desc=Explicitly cites issue 63956 as the known bug identifier.
    - recommend_fixed_version_upgrade earned=0/3 desc=Recommends upgrading to a release containing the fix and explains the version relationship.
- cpu-spike-during-incident-window: verdict=fail failure=diagnosis_critical_gate_failed subtype=- score=50 duration=4m36s workspace=ws-eval-1778484091473-fee04d31
  token usage: input=10704 output=1380 cache_read=36224 cache_write=0 cached_total=36224 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: fail score=0/50
  diagnosis: fail score=0/50
    - identify_capacity_shift_as_primary_driver earned=0/3 desc=Identifies a capacity or node-count shift as the primary driver of the CPU increase.
    - recommend_capacity_adjustment_over_cache_tuning earned=0/3 desc=Recommends capacity adjustment rather than plan cache tuning.
- expensive-query-alert-correlation-during-incident-window: verdict=pass failure=- subtype=- score=92 duration=4m10s workspace=ws-eval-1778484959089-9f11779d
  diagnosis penalties: reject_background_noise_as_primary_root_cause:-3
  token usage: input=2474 output=1580 cache_read=44416 cache_write=0 cached_total=44416 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: pass score=42/50
- expensive-query-spike-during-incident-window: verdict=fail failure=diagnosis_critical_gate_failed subtype=- score=69 duration=4m02s workspace=ws-eval-1778485209312-cccaae73
  token usage: input=1251 output=1465 cache_read=62848 cache_write=0 cached_total=62848 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: fail score=33/50
  diagnosis: fail score=33/50
    - identify_large_scan_or_missing_access_path earned=0/3 desc=Identifies an ineffective access path such as large scan or missing index as the main query bottleneck.
- panic-in-recoverable-goroutine: verdict=pass failure=- subtype=- score=88 duration=5m00s workspace=ws-eval-1778486019543-7351cc3c
  token usage: input=2227 output=1544 cache_read=68480 cache_write=0 cached_total=68480 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: pass score=38/50
- service-degradation-during-incident-window: verdict=pass failure=- subtype=- score=71 duration=4m00s workspace=ws-eval-1778485779958-a171cd48
  token usage: input=2225 output=1654 cache_read=45952 cache_write=0 cached_total=45952 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: pass score=21/50
- sql-tail-latency-spike-during-incident-window: verdict=fail failure=critical_evidence_partial subtype=- score=48 duration=5m29s workspace=ws-eval-1778485451230-a73e4c96
  token usage: input=998 output=1150 cache_read=58752 cache_write=0 cached_total=58752 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: partial score=20/30
    evidence routing: pass score=5/5
    diagnosis: fail score=8/50
  critical evidence: partial score=20/30
    - known_issue_correlation_context_present earned=0/8 desc=Evidence correlates the abnormal heap path with an upstream issue or fixed release.
  diagnosis: fail score=8/50
    - identify_heap_profile_dominant_stack earned=0/3 desc=Uses heap-profile stack evidence to localize the dominant problematic path.
    - mention_github_issue_id earned=0/3 desc=Mentions the concrete GitHub issue id that matches the diagnosed path.
    - recommend_fixed_version_upgrade earned=0/4 desc=Recommends upgrading to a fixed version or build.
- sustained-cpu-pressure-during-incident-window: verdict=fail failure=critical_evidence_partial subtype=- score=48 duration=7m36s workspace=ws-eval-1778491681212-31f560ac
  token usage: input=1237 output=1299 cache_read=52608 cache_write=0 cached_total=52608 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: partial score=16/30
    evidence routing: pass score=5/5
    diagnosis: fail score=12/50
  critical evidence: partial score=16/30
    - allocation_pressure_signal_present earned=0/8 desc=Evidence shows allocation or GC pressure aligned with the CPU spike.
    - heap_profile_hotspot_present earned=0/0 desc=Optional heap profile evidence shows a hotspot aligned with the allocation-heavy path.
  diagnosis: fail score=12/50
    - identify_batch_delete_path_as_primary_driver earned=0/3 desc=Identifies a high-frequency delete path as the primary driver rather than stopping at a generic hot query summary.
    - correlate_log_context_with_primary_workload earned=0/3 desc=Correlates the dominant delete path with log-derived workload context instead of a generic query summary.
    - identify_large_in_list_gc_pressure_chain earned=0/3 desc=Identifies a large IN-list style path as the mechanism linking batch deletes to GC pressure.
- system-unavailability-during-incident-window: verdict=fail failure=critical_evidence_partial subtype=- score=39 duration=4m02s workspace=ws-eval-1778484716929-c6d22bd0
  token usage: input=5434 output=1189 cache_read=62336 cache_write=0 cached_total=62336 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: partial score=11/30
    evidence routing: pass score=5/5
    diagnosis: fail score=8/50
  critical evidence: partial score=11/30
    - capacity_loss_signal_present earned=0/8 desc=Evidence shows a major drop in available coprocessor worker capacity during the incident. snippet=coprocessor-worker
    - restart_or_oom_signal_present earned=0/8 desc=Evidence shows the remaining worker entered restart or memory-failure behavior.
  diagnosis: fail score=8/50
    - identify_worker_capacity_collapse_as_primary_path earned=0/3 desc=Identifies coprocessor worker capacity collapse rather than a pure query-side explanation.
    - identify_restart_loop_as_key_amplifier earned=0/3 desc=Identifies restart or OOM looping on the remaining worker as the main amplifier.
    - recommend_capacity_protection earned=0/4 desc=Recommends capacity-protection actions such as minimum replicas or better scaling guardrails.
    - recommend_worker_health_followup earned=0/4 desc=Recommends follow-up on the initial worker disappearance cause.
- tail-latency-surge-during-incident-window: verdict=pass failure=- subtype=- score=83 duration=3m45s workspace=ws-eval-1778486319562-371b45dd
  token usage: input=444 output=1091 cache_read=45440 cache_write=0 cached_total=45440 cost=$0.0000
  current buckets:
    contract: pass
    execution contract: pass
    workflow: pass score=10/10
    evidence collection: pass score=5/5
    critical evidence: pass score=30/30
    evidence routing: pass score=5/5
    diagnosis: pass score=33/50