Merge branch 'dev' into stage

apps/server/src/log-service-stream.test.ts

@@ -0,0 +1,54 @@
+import { beforeEach, describe, expect, it } from "vitest";
+import { appLogs, type Db } from "@parking/db";
+import { createTestDb } from "@parking/db/testing";
+import { LogService, pinoDbStream } from "./log-service.js";
+
+// pinoDbStream feeds backend warn+ lines into app_logs. Since 2026-07-04 the logger
+// emits level NAMES ("warn") instead of pino's numeric codes (40) — for human-readable
+// container logs — and the stream must accept BOTH encodings (numeric covers any
+// default-configured pino). A level the tee can't resolve falls back to info → not
+// persisted, never a crash.
+
+let db: Db;
+let stream: { write: (line: string) => void };
+let teed: string[];
+
+beforeEach(() => {
+  ({ db } = createTestDb());
+  teed = [];
+  stream = pinoDbStream(new LogService(db), {
+    write: (line: string) => {
+      teed.push(line);
+      return true;
+    },
+  } as unknown as NodeJS.WritableStream);
+});
+
+const rows = () => db.select().from(appLogs).all();
+
+describe("pinoDbStream level encodings", () => {
+  it("persists a LABEL-level warn line (the current logger format)", () => {
+    stream.write(`{"level":"warn","time":"2026-07-04T18:14:11.453Z","msg":"label warn"}\n`);
+    expect(rows()).toHaveLength(1);
+    expect(rows()[0]).toMatchObject({ level: "warn", source: "backend", message: "label warn" });
+  });
+
+  it("still persists a NUMERIC-level error line (legacy/default pino)", () => {
+    stream.write(`{"level":50,"time":1783179038453,"msg":"numeric error"}\n`);
+    expect(rows()[0]).toMatchObject({ level: "error", message: "numeric error" });
+  });
+
+  it("info stays stdout-only in both encodings (teed, not persisted)", () => {
+    stream.write(`{"level":"info","msg":"label info"}\n`);
+    stream.write(`{"level":30,"msg":"numeric info"}\n`);
+    expect(rows()).toHaveLength(0);
+    expect(teed).toHaveLength(2); // stdout tee always happens
+  });
+
+  it("an unresolvable level falls back to info (dropped), never throws", () => {
+    stream.write(`{"level":"loud","msg":"weird"}\n`);
+    stream.write(`not json at all\n`);
+    expect(rows()).toHaveLength(0);
+    expect(teed).toHaveLength(2);
+  });
+});

apps/server/src/log-service.ts

@@ -33,7 +33,10 @@ export interface LogRetention {
 }
 
 export const DEFAULT_RETENTION: LogRetention = {
-  maxAgeDays: Number(process.env.LOG_RETENTION_DAYS ?? 30),
+  // 60 days (~2 months) — the operator's chosen diagnostic window (2026-07-04),
+  // matched by the container-log rotation caps in docker-compose.prod.yml. The row
+  // cap below still bounds a burst regardless of age.
+  maxAgeDays: Number(process.env.LOG_RETENTION_DAYS ?? 60),
   maxRows: Number(process.env.LOG_RETENTION_MAX_ROWS ?? 50_000),
 };
 
@@ -192,9 +195,10 @@ export class LogService {
 
 /**
  * A pino-compatible write stream that forwards BACKEND warn+ lines into the LogService.
- * Pino writes one JSON object per line to this stream; we parse, map the numeric level
- * to a name, and persist. Returned as `{ write }` so it can be passed as pino's stream.
- * stdout still receives the same line (we tee), so console logging is unchanged.
+ * Pino writes one JSON object per line to this stream; we parse, resolve the level
+ * (name or numeric encoding), and persist. Returned as `{ write }` so it can be passed
+ * as pino's stream. stdout still receives the same line (we tee), so console logging is
+ * unchanged.
  */
 export function pinoDbStream(
   service: LogService,
@@ -218,12 +222,17 @@ export function pinoDbStream(
       }
       try {
         const obj = JSON.parse(line) as {
-          level?: number;
+          level?: number | string;
           msg?: string;
           err?: { stack?: string; message?: string };
           [k: string]: unknown;
         };
-        const level = NUM_TO_LEVEL[obj.level ?? 30] ?? "info";
+        // The logger emits level NAMES (formatters.level in server.ts, for human-
+        // readable container logs); a default pino config emits numbers. Accept both.
+        const level: LogLevel =
+          typeof obj.level === "string" && obj.level in LOG_LEVEL_ORDER
+            ? (obj.level as LogLevel)
+            : NUM_TO_LEVEL[typeof obj.level === "number" ? obj.level : 30] ?? "info";
         if (LOG_LEVEL_ORDER[level] < LOG_LEVEL_ORDER[BACKEND_PERSIST_MIN]) return;
         // Strip pino's noisy standard fields from the persisted context.
         const { level: _l, time: _t, pid: _p, hostname: _h, msg, ...rest } = obj;

apps/server/src/server.ts

@@ -71,7 +71,14 @@ export async function buildServer(opts: BuildOptions = {}): Promise<FastifyInsta
   const logService = new LogService(db);
   const app = Fastify({
     logger: {
+      // Level knob: trace|debug|info|warn|error|fatal (pino). Default info; a booth
+      // being diagnosed can run LOG_LEVEL=debug without a code change.
       level: process.env.LOG_LEVEL ?? "info",
+      // Container logs are read by humans (`docker logs` / Komodo), so stamp
+      // ISO-8601 UTC instead of pino's epoch-ms, and level NAMES instead of the
+      // numeric codes (30/40/50). pinoDbStream accepts both encodings.
+      timestamp: () => `,"time":"${new Date().toISOString()}"`,
+      formatters: { level: (label) => ({ level: label }) },
       stream: pinoDbStream(logService, process.stdout),
     },
   });

docker-compose.prod.yml

@@ -26,11 +26,12 @@ services:
       - caddy-config:/config
     depends_on:
       - server
+    # ≈2 months (see the server note): Caddy logs errors only — 10 MB × 5 is plenty.
     logging:
       driver: json-file
       options:
         max-size: "10m"
-        max-file: "3"
+        max-file: "5"
 
   server:
     restart: always
@@ -73,11 +74,17 @@ services:
       - /dev/usb:/dev/usb
     device_cgroup_rules:
       - "c 180:* rmw"
+    # LOG ROTATION (2026-07-04). Docker's json-file driver rotates by SIZE, not time —
+    # these caps are sized to hold ≈2 MONTHS at observed booth rates (the operator's
+    # chosen diagnostic window; revisit if `docker logs` shows less than ~60 days of
+    # history). Server gets the most (request + device chatter): 20 MB × 30 = 600 MB
+    # ceiling. NB: `docker logs` only reaches back as far as these files. The queryable
+    # warn+ store (app_logs, /setup/logs) has its own matching 60-day retention.
     logging:
       driver: json-file
       options:
-        max-size: "10m"
-        max-file: "3"
+        max-size: "20m"
+        max-file: "30"
 
   vision:
     restart: always
@@ -89,11 +96,12 @@ services:
       VISION_RECOGNIZER: fast_alpr
     ports:
       - "127.0.0.1:8089:8089"
+    # ≈2 months (see the server note): vision logs less — 20 MB × 10 = 200 MB ceiling.
     logging:
       driver: json-file
       options:
-        max-size: "10m"
-        max-file: "3"
+        max-size: "20m"
+        max-file: "10"
 
 volumes:
   caddy-data:

wiki/concepts/app-logs.md

@@ -2,7 +2,7 @@
 type: concept
 tags: [parking, observability, diagnostics, logging, frontend, backend]
 sources: []
-updated: 2026-06-19
+updated: 2026-07-04
 status: open
 ---
 
@@ -66,9 +66,26 @@ column — the failed request, error name, component stack, anything), plus pull
 ## Retention (offline appliance ⇒ must be bounded)
 
 Pruned by **age AND a row cap** (a burst could blow past an age-only window): delete older than
-`LOG_RETENTION_DAYS` (default 30) **and** keep only the newest `LOG_RETENTION_MAX_ROWS` (default
-50 000). Runs **hourly** (unref'd timer) + once at startup. Both env-configurable. Same "prunable,
-not precious" durability class as `device_events` — the opposite of the append-only ledger.
+`LOG_RETENTION_DAYS` (default **60** — the operator's ≈2-month diagnostic window, 2026-07-04;
+was 30) **and** keep only the newest `LOG_RETENTION_MAX_ROWS` (default 50 000). Runs **hourly**
+(unref'd timer) + once at startup. Both env-configurable. Same "prunable, not precious"
+durability class as `device_events` — the opposite of the append-only ledger.
+
+## Container (stdout) logs — the OTHER log store (2026-07-04)
+
+`docker logs` is a separate, size-bounded store from `app_logs` — it holds **everything**
+(info/debug too), while `app_logs` keeps only warn+. Three knobs, all set 2026-07-04:
+
+- **Rotation:** Docker's json-file driver rotates by SIZE, not time; the caps in
+  `docker-compose.prod.yml` are sized to hold **≈2 months** at observed booth rates (server
+  20 MB × 30, vision 20 MB × 10, proxy 10 MB × 5). `docker logs` reaches back only that far —
+  revisit the caps if it shows under ~60 days. (Dev compose is uncapped — laptop concern only.)
+- **Human-readable lines:** the pino logger stamps **ISO-8601 UTC** `time` (was epoch-ms) and
+  **level NAMES** (`"warn"`, was `40`) via `timestamp` + `formatters.level` in `server.ts`.
+  `pinoDbStream` accepts BOTH level encodings, so the app_logs tee survives either config.
+- **Level knob:** `LOG_LEVEL` env (trace|debug|info|warn|error|fatal; default `info`) — a booth
+  under diagnosis runs `LOG_LEVEL=debug` with no code change; warn+ persistence is unaffected
+  (it filters independently in the tee).
 
 ## The booth viewer
 

wiki/log.md

@@ -2272,3 +2272,13 @@ card, unknown SUB- code) still signs the normal anomaly; enrolled credentials ma
 filter and can never be hidden by it. Works for legacy unprefixed reads too, so the feed cleans up
 before the vendor-tool visit. 6 new tests; suite 278 green. Also this session: reader channel
 tagging (clone defense) — see the prior entry.
+
+## [2026-07-04] update | Logging: 2-month rotation, ISO timestamps, level names
+
+Operator asked for bounded container logs (~2 months), human-readable timestamps, and clarity on
+levels. Findings + changes on [[app-logs]]: levels EXISTED (LOG_LEVEL env, pino, warn+ teed to
+app_logs); the "level":30 numbers and epoch-ms times were pino defaults — the logger now stamps
+ISO-8601 UTC + level names (pinoDbStream hardened to accept both encodings so the DB tee can't
+silently break). Rotation: docker json-file caps in docker-compose.prod.yml resized from 10m×3
+(≈30 MB!) to ≈2 months by volume (server 20m×30, vision 20m×10, proxy 10m×5; json-file rotates by
+SIZE — time-based isn't a driver feature). app_logs retention default aligned 30→60 days.