From 8675c1f7c3eedbb287c85de690341b7208ab5489 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ullrich=20Sch=C3=A4fer?= <hi+github@ullrich.is>
Date: Sun, 24 May 2026 12:37:09 +0200
Subject: [PATCH] fix(journal): reuse a dedicated pool for /api/health instead
 of per-call connect
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The previous handler opened a fresh postgres client (max: 1) on every
call to /api/health and tore it down in the finally block. Under the
prod monitoring cadence (probes every few seconds), that's a fresh
TCP + TLS + auth handshake on every probe, plus connection-table
churn on the Postgres side — fine for the trickle of curl-ish manual
checks, slow-bleed under blackbox monitoring.

Now we cache a module-level singleton postgres client dedicated to
/api/health (max: 2, idle_timeout: 30) and reuse it across calls.
Separate from the app's main DB pool (via @trails-cool/db's createDb)
on purpose — so a starvation event on the main pool doesn't fail the
liveness check and trigger a restart loop.

Full repo: pnpm typecheck, pnpm lint, pnpm test all green.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
---
 apps/journal/server.ts | 20 +++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

diff --git a/apps/journal/server.ts b/apps/journal/server.ts
index 89dbf93..90c3c39 100644
--- a/apps/journal/server.ts
+++ b/apps/journal/server.ts
@@ -9,7 +9,7 @@ import { randomUUID } from "node:crypto";
 import { httpRequestDuration, registry } from "./app/lib/metrics.server.ts";
 import { createBoss, startWorker } from "@trails-cool/jobs";
 import { getDatabaseUrl } from "@trails-cool/db";
-import postgres from "postgres";
+import postgres, { type Sql } from "postgres";
 
 // Sentry DSN is read from env so self-hosted instances don't ship their
 // errors to the trails.cool flagship Sentry by default. The flagship
@@ -76,17 +76,27 @@ async function handleMetrics(_req: IncomingMessage, res: ServerResponse): Promis
 
 const version = process.env.SENTRY_RELEASE ?? "dev";
 
+// Module-level singleton postgres client dedicated to /api/health. The
+// previous handler opened a fresh client + connection on every call,
+// which OOM'd the process under monitoring load (probes hit /api/health
+// every few seconds). `max: 2` is plenty for liveness checks; the main
+// app DB pool is separate (via @trails-cool/db's createDb).
+let healthClient: Sql | null = null;
+function getHealthClient(): Sql {
+  if (!healthClient) {
+    healthClient = postgres(getDatabaseUrl(), { max: 2, idle_timeout: 30 });
+  }
+  return healthClient;
+}
+
 async function handleHealth(_req: IncomingMessage, res: ServerResponse): Promise<void> {
-  const client = postgres(getDatabaseUrl(), { max: 1 });
   try {
-    await client`SELECT 1`;
+    await getHealthClient()`SELECT 1`;
     res.writeHead(200, { "Content-Type": "application/json" });
     res.end(JSON.stringify({ status: "ok", version, db: "connected" }));
   } catch {
     res.writeHead(503, { "Content-Type": "application/json" });
     res.end(JSON.stringify({ status: "degraded", version, db: "unreachable" }));
-  } finally {
-    await client.end();
   }
 }