Prepare KarlDNS production deployment

karldns/store/metrics.py

@@ -0,0 +1,84 @@
+from __future__ import annotations
+
+from datetime import UTC, datetime
+
+
+class MetricsMixin:
+    def increment_metric(self, name: str, amount: float = 1.0) -> None:
+        now = datetime.now(UTC).isoformat()
+        with self.connect() as conn:
+            conn.execute(
+                """
+                INSERT INTO service_metrics(name, value, updated_at)
+                VALUES (?, ?, ?)
+                ON CONFLICT(name) DO UPDATE SET
+                  value=value + excluded.value,
+                  updated_at=excluded.updated_at
+                """,
+                (name, amount, now),
+            )
+
+    def set_metric(self, name: str, value: float) -> None:
+        now = datetime.now(UTC).isoformat()
+        with self.connect() as conn:
+            conn.execute(
+                """
+                INSERT INTO service_metrics(name, value, updated_at)
+                VALUES (?, ?, ?)
+                ON CONFLICT(name) DO UPDATE SET
+                  value=excluded.value,
+                  updated_at=excluded.updated_at
+                """,
+                (name, value, now),
+            )
+
+    def get_service_metrics(self) -> dict[str, float]:
+        with self.connect() as conn:
+            rows = conn.execute("SELECT name, value FROM service_metrics").fetchall()
+        return {str(row["name"]): float(row["value"]) for row in rows}
+
+    def record_update_metrics(
+        self,
+        latency_ms: float,
+        *,
+        is_success: bool = False,
+        is_dns_failure: bool = False,
+        is_reconcile_failure: bool = False,
+        is_bad_ip: bool = False,
+    ) -> None:
+        increments = {
+            "update_requests_total": 1,
+            "update_latency_ms_total": latency_ms,
+            "update_latency_samples_total": 1,
+        }
+        if is_success:
+            increments["update_success_total"] = 1
+        if is_dns_failure:
+            increments["dns_failures_total"] = 1
+        if is_reconcile_failure:
+            increments["dns_reconcile_failures_total"] = 1
+        if is_bad_ip:
+            increments["bad_ip_total"] = 1
+
+        now = datetime.now(UTC).isoformat()
+        with self.connect() as conn:
+            conn.executemany(
+                """
+                INSERT INTO service_metrics(name, value, updated_at)
+                VALUES (?, ?, ?)
+                ON CONFLICT(name) DO UPDATE SET
+                  value=value + excluded.value,
+                  updated_at=excluded.updated_at
+                """,
+                [(name, value, now) for name, value in increments.items()],
+            )
+            conn.execute(
+                """
+                INSERT INTO service_metrics(name, value, updated_at)
+                VALUES ('last_update_latency_ms', ?, ?)
+                ON CONFLICT(name) DO UPDATE SET
+                  value=excluded.value,
+                  updated_at=excluded.updated_at
+                """,
+                (latency_ms, now),
+            )