Add wallclock/duration delta to per-PR journey-delta comment

Extend JourneyDelta with DurationMSDelta, computed in ComputeJourneyDeltas
from the existing journeymetrics.Record.DurationMS field, and render it as
a new Duration Δ (ms) column in RenderJourneyDeltaComment.

Author: Spike Test

internal/release/journeydelta.go

@@ -62,11 +62,12 @@ type JourneyDelta struct {
 	ScenarioID     string
 	Runtime        string
 	Model          string
-	HasBaseline    bool
-	BaselineRunURL string
-	TurnsDelta     int
-	TokensDelta    journeymetrics.TokenTotals
-	CostDeltaUSD   float64
+	HasBaseline     bool
+	BaselineRunURL  string
+	TurnsDelta      int
+	DurationMSDelta int64
+	TokensDelta     journeymetrics.TokenTotals
+	CostDeltaUSD    float64
 }
 
 // ComputeJourneyDeltas computes, for every scenario/runtime/model the current
@@ -94,6 +95,7 @@ func ComputeJourneyDeltas(baseline journeymetrics.Ledger, current []journeymetri
 			d.BaselineRunURL = base.RunURL
 		}
 		d.TurnsDelta = cur.Turns - base.Turns
+		d.DurationMSDelta = cur.DurationMS - base.DurationMS
 		d.TokensDelta = journeymetrics.TokenTotals{
 			Input:         cur.Tokens.Input - base.Tokens.Input,
 			Output:        cur.Tokens.Output - base.Tokens.Output,
@@ -138,20 +140,20 @@ func RenderJourneyDeltaComment(deltas []JourneyDelta) string {
 		b.WriteString("No journey metrics observations were produced by this run.\n")
 		return b.String()
 	}
-	b.WriteString("| Scenario | Runtime | Model | Turns Δ | Cache Read Δ | Cache Creation Δ | Tokens Δ (total) | Cost Δ (USD) | Baseline |\n")
-	b.WriteString("| --- | --- | --- | --- | --- | --- | --- | --- | --- |\n")
+	b.WriteString("| Scenario | Runtime | Model | Turns Δ | Cache Read Δ | Cache Creation Δ | Tokens Δ (total) | Cost Δ (USD) | Duration Δ (ms) | Baseline |\n")
+	b.WriteString("| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |\n")
 	for _, d := range deltas {
 		if !d.HasBaseline {
-			fmt.Fprintf(&b, "| %s | %s | %s | n/a (new) | n/a (new) | n/a (new) | n/a (new) | n/a (new) | none (new observation) |\n",
+			fmt.Fprintf(&b, "| %s | %s | %s | n/a (new) | n/a (new) | n/a (new) | n/a (new) | n/a (new) | n/a (new) | none (new observation) |\n",
 				d.ScenarioID, d.Runtime, d.Model)
 			continue
 		}
 		baseline := "latest published"
 		if d.BaselineRunURL != "" {
 			baseline = fmt.Sprintf("[latest published](%s)", d.BaselineRunURL)
 		}
-		fmt.Fprintf(&b, "| %s | %s | %s | %+d | %+d | %+d | %+d | %+.4f | %s |\n",
-			d.ScenarioID, d.Runtime, d.Model, d.TurnsDelta, d.TokensDelta.CacheRead, d.TokensDelta.CacheCreation, d.TokensDelta.Total, d.CostDeltaUSD, baseline)
+		fmt.Fprintf(&b, "| %s | %s | %s | %+d | %+d | %+d | %+d | %+.4f | %+d | %s |\n",
+			d.ScenarioID, d.Runtime, d.Model, d.TurnsDelta, d.TokensDelta.CacheRead, d.TokensDelta.CacheCreation, d.TokensDelta.Total, d.CostDeltaUSD, d.DurationMSDelta, baseline)
 	}
 	return b.String()
 }

internal/release/journeydelta_test.go

@@ -200,9 +200,9 @@ func TestRenderJourneyDeltaCommentShowsTokenClassBreakdown(t *testing.T) {
 		t.Fatalf("comment body missing the data row:\n%s", body)
 	}
 	cells := strings.Split(row, "|")
-	// | (empty) | Scenario | Runtime | Model | Turns Δ | Cache Read Δ | Cache Creation Δ | Tokens Δ (total) | Cost Δ (USD) | Baseline | (empty) |
-	if len(cells) != 11 {
-		t.Fatalf("data row has %d cells, want 11 (header shape changed?): %q", len(cells), row)
+	// | (empty) | Scenario | Runtime | Model | Turns Δ | Cache Read Δ | Cache Creation Δ | Tokens Δ (total) | Cost Δ (USD) | Duration Δ (ms) | Baseline | (empty) |
+	if len(cells) != 12 {
+		t.Fatalf("data row has %d cells, want 12 (header shape changed?): %q", len(cells), row)
 	}
 	cacheReadCell := strings.TrimSpace(cells[5])
 	cacheCreationCell := strings.TrimSpace(cells[6])
@@ -218,6 +218,57 @@ func TestRenderJourneyDeltaCommentShowsTokenClassBreakdown(t *testing.T) {
 	}
 }
 
+// TestJourneyDeltaCommentShowsDurationDelta is the TDD proof for the
+// wallclock/duration delta: computes deltas from a baseline ledger and a
+// current PR run with distinct DurationMS values, then asserts both the
+// computed struct field and the rendered comment's Duration Δ column show the
+// exact millisecond delta (current - baseline), not a total or a placeholder.
+func TestJourneyDeltaCommentShowsDurationDelta(t *testing.T) {
+	baseline := journeymetrics.Ledger{
+		Scenarios: []journeymetrics.ScenarioLedgerEntry{
+			{
+				ScenarioID: "shallow-boot-window",
+				Observations: []journeymetrics.Record{
+					{
+						ScenarioID: "shallow-boot-window", Runtime: "claude", Model: "claude-sonnet-4-6",
+						DurationMS: 12000, CapturedAt: "2026-06-20T00:00:00Z",
+						RunURL: "https://github.com/spacedock-dev/spacedock/actions/runs/27931963802",
+					},
+				},
+			},
+		},
+	}
+	current := []journeymetrics.Record{
+		{
+			ScenarioID: "shallow-boot-window", Runtime: "claude", Model: "claude-sonnet-4-6",
+			DurationMS: 15500,
+		},
+	}
+
+	deltas := ComputeJourneyDeltas(baseline, current)
+	if len(deltas) != 1 {
+		t.Fatalf("deltas = %d, want 1", len(deltas))
+	}
+	if deltas[0].DurationMSDelta != 3500 {
+		t.Fatalf("DurationMSDelta = %d, want 3500 (15500-12000)", deltas[0].DurationMSDelta)
+	}
+
+	body := RenderJourneyDeltaComment(deltas)
+	var row string
+	for _, line := range strings.Split(body, "\n") {
+		if strings.HasPrefix(line, "| shallow-boot-window") {
+			row = line
+			break
+		}
+	}
+	if row == "" {
+		t.Fatalf("comment body missing the data row:\n%s", body)
+	}
+	if !strings.Contains(row, "+3500") {
+		t.Fatalf("data row missing Duration Δ +3500(ms): %q", row)
+	}
+}
+
 // TestRenderJourneyDeltaCommentRendersNoBaselineAsNewNotSelfDelta proves a
 // scenario/model with no matching baseline observation renders "n/a (new)" in
 // its delta cells rather than a self-delta against an implicit zero baseline