feat: add quantiles() method to LatencyTracker for OpenMetrics

Returns {count, sum, 0.5, 0.95, 0.99} in seconds for Prometheus
summary exposition. Companion to the existing stats() (milliseconds).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

tests/test_metrics.py

@@ -52,6 +52,23 @@ class TestLatencyTracker:
         assert s["min"] == 500.0
         assert s["max"] == 500.0
 
+    def test_quantiles_empty(self):
+        lt = LatencyTracker()
+        assert lt.quantiles() is None
+
+    def test_quantiles_seconds(self):
+        lt = LatencyTracker()
+        for i in range(1, 101):
+            lt.record(i / 1000)
+        q = lt.quantiles()
+        assert q is not None
+        assert q["count"] == 100
+        assert 0.050 <= q["0.5"] <= 0.052
+        assert 0.095 <= q["0.95"] <= 0.097
+        assert 0.099 <= q["0.99"] <= 0.101
+        assert "sum" in q
+        assert q["sum"] > 0
+
 
 # -- RateTracker -------------------------------------------------------------