fix(metrics): time-series rows get time-ordered ids, so an insert stops costing a disk read (#5149)

Measured on dev (385a579), agent posts cut to five samples still dropped:
flush=8.8s for 990 rows, 6.6s for 680, 8.4s for 1000 — about 110 rows/s,
~9 ms a row, with rules and evaluation near zero. One random read per
row fits a random uuid4 primary key whose index no longer fits in memory:
every new row lands on a random page of it. (Inferred from the rate; the
table and index sizes have not been read.)

- steward.core.ids.time_ordered_id: a UUIDv7-layout string (48-bit ms
  timestamp, then random), same 36-char shape as the stored uuid4 ids, so
  no migration. Python's uuid7() arrives in 3.14; the image runs 3.13.
- Used for plugin_metrics and its hourly rollup, monitor_results,
  audit_events, the docker metrics/hourly/events/logs tables, and the
  traefik and unifi time series.
- MAX_SAMPLES_PER_INGEST 5 → 2 (~400 rows) for now: it fits the agent's
  10 s even at the old rate, so hosts land data whichever way the new rate
  comes out. Raise it once the new rate is measured.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

plugins/docker/models.py

@@ -1,6 +1,5 @@
 # plugins/docker/models.py
 from __future__ import annotations
-import uuid
 from datetime import datetime, timezone
 from sqlalchemy import (
     BigInteger, Boolean, DateTime, Float, ForeignKey, Index, Integer, String, Text,
@@ -8,6 +7,7 @@ from sqlalchemy import (
 )
 from sqlalchemy.orm import Mapped, mapped_column
 from steward.models.base import Base
+from steward.core.ids import time_ordered_id
 
 
 class DockerContainer(Base):
@@ -67,7 +67,7 @@ class DockerMetric(Base):
     __tablename__ = "docker_metrics"
 
     id: Mapped[str] = mapped_column(
-        String(36), primary_key=True, default=lambda: str(uuid.uuid4())
+        String(36), primary_key=True, default=time_ordered_id
     )
     host_id: Mapped[str] = mapped_column(
         String(36), ForeignKey("hosts.id", ondelete="CASCADE"), nullable=False, index=True
@@ -103,7 +103,7 @@ class DockerMetricHourly(Base):
     __tablename__ = "docker_metrics_hourly"
 
     id: Mapped[str] = mapped_column(
-        String(36), primary_key=True, default=lambda: str(uuid.uuid4())
+        String(36), primary_key=True, default=time_ordered_id
     )
     host_id: Mapped[str] = mapped_column(
         String(36), ForeignKey("hosts.id", ondelete="CASCADE"), nullable=False
@@ -139,7 +139,7 @@ class DockerEvent(Base):
     __tablename__ = "docker_events"
 
     id: Mapped[str] = mapped_column(
-        String(36), primary_key=True, default=lambda: str(uuid.uuid4())
+        String(36), primary_key=True, default=time_ordered_id
     )
     host_id: Mapped[str] = mapped_column(
         String(36), ForeignKey("hosts.id", ondelete="CASCADE"), nullable=False
@@ -176,7 +176,7 @@ class DockerLog(Base):
     __tablename__ = "docker_logs"
 
     id: Mapped[str] = mapped_column(
-        String(36), primary_key=True, default=lambda: str(uuid.uuid4())
+        String(36), primary_key=True, default=time_ordered_id
     )
     host_id: Mapped[str] = mapped_column(
         String(36), ForeignKey("hosts.id", ondelete="CASCADE"), nullable=False, index=True

plugins/host_agent/routes.py

@@ -343,7 +343,9 @@ SLOW_INGEST_SECONDS = 2.0
 # instance that did not finish inside the agent's 10 s, the rollback left the
 # buffer full, and every later post failed the same way. The newest samples
 # are kept: the older ones would be lost to the agent's ring buffer anyway.
-MAX_SAMPLES_PER_INGEST = 5
+# Two (~400 rows) fits 10 s even at the ~110 rows/s measured before
+# plugin_metrics got time-ordered ids; raise it once the new rate is known.
+MAX_SAMPLES_PER_INGEST = 2
 
 
 def newest_samples(samples: list, limit: int) -> tuple[list, int]:

plugins/traefik/models.py

@@ -1,16 +1,16 @@
 # plugins/traefik/models.py
 from __future__ import annotations
-import uuid
 from datetime import datetime, timezone
 from sqlalchemy import DateTime, Float, Integer, String, Text
 from sqlalchemy.orm import Mapped, mapped_column
 from steward.models.base import Base
+from steward.core.ids import time_ordered_id
 
 
 class TraefikMetric(Base):
     __tablename__ = "traefik_metrics"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     router_name: Mapped[str] = mapped_column(String(255), nullable=False)
     scraped_at: Mapped[datetime] = mapped_column(
         DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc)
@@ -41,7 +41,7 @@ class TraefikAccessSummary(Base):
     """Aggregated traffic-origin summary — one row per scheduler tick."""
     __tablename__ = "traefik_access_summaries"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     period_start: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False)
     period_end: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False)
     total_requests: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
@@ -57,7 +57,7 @@ class TraefikGlobalStat(Base):
     """One row per scrape — process-level and config stats."""
     __tablename__ = "traefik_global_stats"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     scraped_at: Mapped[datetime] = mapped_column(
         DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc)
     )

plugins/unifi/models.py

@@ -1,10 +1,10 @@
 # plugins/unifi/models.py
 from __future__ import annotations
-import uuid
 from datetime import datetime, timezone
 from sqlalchemy import Boolean, DateTime, Float, Integer, String, Text
 from sqlalchemy.orm import Mapped, mapped_column
 from steward.models.base import Base
+from steward.core.ids import time_ordered_id
 
 
 # ── Time-series (one row per scrape) ──────────────────────────────────────────
@@ -13,7 +13,7 @@ class UnifiWanStat(Base):
     """WAN interface health — one row per scrape."""
     __tablename__ = "unifi_wan_stats"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     scraped_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc))
     is_up: Mapped[bool] = mapped_column(Boolean, nullable=False, default=False)
     wan_ip: Mapped[str | None] = mapped_column(String(64), nullable=True)
@@ -27,7 +27,7 @@ class UnifiClientSnapshot(Base):
     """Aggregated client counts + top talkers — one row per scrape."""
     __tablename__ = "unifi_client_snapshots"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     scraped_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc))
     total_clients: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
     wireless_clients: Mapped[int] = mapped_column(Integer, nullable=False, default=0)
@@ -41,7 +41,7 @@ class UnifiWlanStat(Base):
     """Per-SSID/radio stats — one row per scrape per SSID+radio combination."""
     __tablename__ = "unifi_wlan_stats"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     scraped_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc))
     ssid: Mapped[str] = mapped_column(String(255), nullable=False)
     radio: Mapped[str | None] = mapped_column(String(8), nullable=True)    # ng=2.4GHz, na=5GHz, 6e=6GHz
@@ -56,7 +56,7 @@ class UnifiDpiSnapshot(Base):
     """DPI application/category traffic snapshot — one row per scrape (JSON blob)."""
     __tablename__ = "unifi_dpi_snapshots"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     scraped_at: Mapped[datetime] = mapped_column(DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc))
     # JSON: [{"cat_id", "cat_name", "rx_bytes", "tx_bytes"}] aggregated across all clients
     by_category_json: Mapped[str] = mapped_column(Text, nullable=False, default="[]")

steward/core/ids.py

@@ -0,0 +1,25 @@
+"""Row ids for tables that grow by time.
+
+A random uuid4 primary key puts every new row on a random page of the key's
+index; once that index outgrows memory, each insert waits for a disk read.
+Measured on the live instance (#5149): plugin_metrics took ~9 ms a row, about
+110 rows/s, so a host agent's post could not finish inside its 10 s timeout.
+A time-ordered id lands every new row at the index's right edge instead.
+"""
+from __future__ import annotations
+
+import os
+import time
+import uuid
+
+
+def time_ordered_id() -> str:
+    """A UUIDv7-layout id (RFC 9562) as a 36-character string: a 48-bit
+    millisecond timestamp, then 74 random bits. Same shape and column type as
+    the uuid4 ids already stored, so no table changes; Python's own uuid7()
+    arrives only in 3.14."""
+    ms = time.time_ns() // 1_000_000
+    value = (ms & 0xFFFF_FFFF_FFFF) << 80 | int.from_bytes(os.urandom(10), "big")
+    value = (value & ~(0xF << 76)) | (0x7 << 76)     # version 7
+    value = (value & ~(0x3 << 62)) | (0x2 << 62)     # RFC 9562 variant
+    return str(uuid.UUID(int=value))

steward/models/audit.py

@@ -1,15 +1,15 @@
 from __future__ import annotations
-import uuid
 from datetime import datetime, timezone
 from sqlalchemy import DateTime, ForeignKey, String, Text
 from sqlalchemy.orm import Mapped, mapped_column
 from .base import Base
+from steward.core.ids import time_ordered_id
 
 
 class AuditEvent(Base):
     __tablename__ = "audit_events"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     timestamp: Mapped[datetime] = mapped_column(
         DateTime(timezone=True), nullable=False, default=lambda: datetime.now(timezone.utc)
     )

steward/models/metrics.py

@@ -1,15 +1,15 @@
 from __future__ import annotations
-import uuid
 from datetime import datetime, timezone
 from sqlalchemy import DateTime, Float, Index, Integer, String, UniqueConstraint
 from sqlalchemy.orm import Mapped, mapped_column
 from .base import Base
+from steward.core.ids import time_ordered_id
 
 
 class PluginMetric(Base):
     __tablename__ = "plugin_metrics"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     source_module: Mapped[str] = mapped_column(String(64), nullable=False)
     resource_name: Mapped[str] = mapped_column(String(255), nullable=False)
     metric_name: Mapped[str] = mapped_column(String(128), nullable=False)
@@ -40,7 +40,7 @@ class PluginMetricHourly(Base):
     """
     __tablename__ = "plugin_metrics_hourly"
 
-    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=lambda: str(uuid.uuid4()))
+    id: Mapped[str] = mapped_column(String(36), primary_key=True, default=time_ordered_id)
     source_module: Mapped[str] = mapped_column(String(64), nullable=False)
     resource_name: Mapped[str] = mapped_column(String(255), nullable=False)
     metric_name: Mapped[str] = mapped_column(String(128), nullable=False)

steward/models/monitors.py

@@ -8,6 +8,7 @@ from sqlalchemy import (
 )
 from sqlalchemy.orm import Mapped, mapped_column
 from .base import Base
+from steward.core.ids import time_ordered_id
 
 
 class MonitorType(str, enum.Enum):
@@ -93,7 +94,7 @@ class MonitorResult(Base):
     __tablename__ = "monitor_results"
 
     id: Mapped[str] = mapped_column(
-        String(36), primary_key=True, default=lambda: str(uuid.uuid4())
+        String(36), primary_key=True, default=time_ordered_id
     )
     monitor_id: Mapped[str] = mapped_column(
         String(36), ForeignKey("monitors.id", ondelete="CASCADE"),

tests/core/test_ids.py

@@ -0,0 +1,34 @@
+"""Time-series rows get time-ordered ids, so a new row lands at the right edge
+of the primary-key index rather than on a random page of it (#5149)."""
+import uuid
+
+from steward.core import ids
+from steward.core.ids import time_ordered_id
+from steward.models.metrics import PluginMetric
+
+
+def test_the_id_is_a_version_7_uuid_string():
+    raw = time_ordered_id()
+    u = uuid.UUID(raw)
+    assert len(raw) == 36 and str(u) == raw
+    assert u.version == 7 and u.variant == uuid.RFC_4122
+
+
+def test_ids_sort_by_the_millisecond_they_were_made(monkeypatch):
+    made = []
+    for ms in (1_791_327_000_000, 1_791_327_000_001, 1_791_327_060_000):
+        monkeypatch.setattr(ids.time, "time_ns", lambda ms=ms: ms * 1_000_000)
+        made.append(time_ordered_id())
+    assert made == sorted(made)
+    assert len({m[:13] for m in made}) == 3     # the timestamp leads the string
+
+
+def test_two_ids_in_one_millisecond_differ(monkeypatch):
+    monkeypatch.setattr(ids.time, "time_ns", lambda: 1_791_327_000_000 * 1_000_000)
+    assert time_ordered_id() != time_ordered_id()
+
+
+def test_plugin_metrics_use_it():
+    # SQLAlchemy wraps a no-argument default as fn(context); call it as it would.
+    made = PluginMetric.__table__.c.id.default.arg(None)
+    assert uuid.UUID(made).version == 7