IoT telemetry on Next.js 16 (App Router), Postgres (Neon) and Better Auth

Next.js 16 App Router: file-based routing, server components, and the Edge proxy (Next 16's renamed middleware).

Postgres on Neon via Drizzle ORM and the postgres-js driver.

Better Auth: self-hosted auth running inside your app against your Postgres (Drizzle adapter).

IoT telemetry: user-owned devices, append-only sensor readings, threshold alert rules, and fired alert instances.

Database client

import { drizzle } from "drizzle-orm/postgres-js";
import postgres from "postgres";

// Neon pooled endpoint = PgBouncer transaction mode → prepared statements off.
// ponytail: single module-level client; the serverless runtime + PgBouncer do
// the pooling, so no custom pool/globalThis singleton dance needed.
const client = postgres(process.env.DATABASE_URL!, { prepare: false });

export const db = drizzle({ client });

user-owned device registry with a unique device_key and an online/offline status CHECK

append-only rows of metric + numeric value per device, indexed for time-range rollups

per-device threshold rules (gt/lt/gte/lte comparator CHECK) and the alert instances they fire, with firing/resolved lifecycle

Auth runs in proxy.ts (Next 16's renamed middleware) on the Edge runtime: it gates on the session cookie's presence only — full session validation happens in Server Components and route handlers, not in the proxy.

prepare: false is mandatory — Neon's pooled endpoint is PgBouncer in transaction mode, where server-side prepared statements break across the pool.

Drizzle is paired here (not Prisma): Prisma's prepared-statement reliance is incompatible with transaction-mode pooling.

Self-hosted: Better Auth owns the user/session/account/verification tables. This stack emits them (db/auth-schema.ts) and hands them to the Drizzle adapter, so app-type schemas can foreign-key `user` directly.

devices.device_key carries a unique constraint — it is the physical device's identity token and must be generated once at provisioning time, never regenerated.

sensor_readings is append-only (no update path, composite index on device_id + recorded_at) — roll up by device + time window for dashboards rather than mutating any running aggregate.

IoT telemetry

The device stamps the time, not the server: sensor_readings.recorded_at is notNull with no default, so whatever the firmware puts on the wire is what lands. A gateway buffering through an outage therefore replays honestly, its backlog keeping true instants instead of collapsing onto the moment it reconnected — and readings arrive out of order as a matter of course, so the latest reading for a device means ordering by recorded_at, never taking a max of anything else. Readings are a narrow long table: device_id, a free-text metric, a value, and that timestamp. On Postgres value is numeric, which is arbitrary precision; on MySQL it is decimal(20,6), because a bare mysql decimal is (10,0) and would silently round every temperature to a whole degree.

Six fractional digits is the ceiling on that side. Devices are the scoping root. device_key is UNIQUE and serves as the hardware's identity at ingest — issued once at provisioning, and the only credential this schema knows about. owner_id foreign-keys Better Auth's user, with idx_device_owner behind the fleet list. status is a stored flag held to 'online' or 'offline' by devices_status_check; it is not derived from readings, so a heartbeat job or the ingest handler has to write it, and it stays wrong until something does. The alerting half is two tables rather than one because a declaration and a firing have different lifetimes.

alert_rules is a per-device threshold — a metric, a comparator held by alert_rules_comparator_check to gt, lt, gte or lte, and a threshold in the same numeric type as value — and nothing stops several rules covering one metric. alerts is the instance: status guarded to 'firing' or 'resolved', triggered_at defaulted, resolved_at nullable until something closes it. It stores both rule_id and device_id even though the device is reachable through the rule, and that redundancy is the point — idx_alert_device answers what is firing on this device with no join. Cascades run the whole way down from devices, so decommissioning one deletes its readings, its rules and its alert history in a single statement. If the telemetry has to outlive the hardware, keep the device row and stop writing to it.

Next.js 16 (App Router)

Next 16's App Router keeps the whole application under src/: the scaffold runs create-next-app with --src-dir and --import-alias @/*, so `@/` resolves to src/* and the verified files land on top of a stock project rather than replacing it. initCode writes the database client to src/lib/db.ts, then appends the auth fragment's own files — identity tables at src/db/auth-schema.ts, the auth instance at src/lib/auth.ts, a catch-all handler under src/app/api/auth/, and src/proxy.ts. Each file has exactly one owner; the framework never re-emits the ones auth brought. Server code has two shapes here and they are not interchangeable. A Server Component runs on the server and imports { db } from "@/lib/db" directly, so a page can await a Drizzle query with no API route in between.

Anything that needs a URL — an OAuth callback, a Polar or Clerk webhook, a mutation posted from the client — is a route handler at src/app/api/<path>/route.ts exporting GET or POST. Session checking is split across two tiers on purpose. src/proxy.ts (Next 16's rename of middleware.ts; under --src-dir it must sit beside src/app or Next silently ignores it) runs on the Edge runtime with a config.matcher listing the guarded prefixes — /dashboard/:path* and /settings/:path* out of the box. It only asks whether a session cookie exists and redirects to /sign-in when it does not: no database round trip at the edge. The authoritative check is auth.api.getSession() inside the Server Component or route handler that actually reads rows.

What that means when you build on it: widening the protected surface is a one-line change to the matcher array, but the proxy is not the security boundary — a request carrying any session cookie reaches the page, and the page decides. Keep the real check next to the data. The UI overlay follows the same split, with auth screens under src/app/(auth)/ and the shell and dashboard under src/app/(app)/.

Postgres (Neon)

Postgres here is Neon reached through postgres-js, with Drizzle's pg-core dialect on top: drizzle({ client }) over a single module-level postgres(DATABASE_URL, { prepare: false }). That flag is not a preference. Neon's pooled (-pooler) endpoint is PgBouncer in transaction mode, where a backend is handed to a different session between statements, so server-side prepared statements break across the pool — and the same constraint is why this axis pairs with Drizzle rather than Prisma. One client per module is enough: PgBouncer and the runtime do the pooling, so there is no globalThis singleton dance. The schemas built on this dialect make three recurring type decisions. Primary keys are uuid(...).primaryKey().defaultRandom(), so ids come from the database. Timestamps are timestamp(..., { withTimezone: true }).defaultNow() — timestamptz, an absolute instant.

Closed value sets are text plus a CHECK constraint rather than pgEnum, so shipping a new role or subscription status is an ordinary constraint change instead of an ALTER TYPE migration. Counters are bigint({ mode: "number" }), and Better Auth's text user.id is referenced as text by the app tables rather than recast. Operationally, transaction-mode pooling forbids anything that spans statements on one backend: LISTEN/NOTIFY, session-scoped SET, advisory-lock sessions, WITH HOLD cursors. Those paths use Neon's direct endpoint instead. The connection client also changes with the deploy target — max: 1 per short-lived serverless instance, a real reused pool (max 10, idle_timeout 20) in a long-running Node process, and on Cloudflare Workers postgres-js is replaced outright by @neondatabase/serverless over HTTP, because Workers have no TCP sockets.

The capability that exists only on this side of the matrix is row-level security. Multi-tenant schemas ship ENABLE plus FORCE ROW LEVEL SECURITY with policies keyed on current_setting('app.current_org_id', true), which withTenant() sets per transaction — unset context yields no rows, so isolation fails closed inside the database rather than in application code. It requires a dedicated NOBYPASSRLS role: Neon's default neondb_owner carries BYPASSRLS, and connecting as it makes every policy silently inert.

Better Auth

Better Auth is a TypeScript library, not a service. The process that serves your pages is the process that hashes passwords and issues sessions, and the session rows sit in the same database as your application data. This fragment authors the four tables Better Auth expects — user, session, account, verification — into db/auth-schema.ts and passes that module to drizzleAdapter(db, { provider, schema: authSchema }) inside lib/auth.ts. Better Auth never creates tables at runtime; its CLI normally generates them, and authoring them here means the identity schema goes through the same migration proof as the app-type tables.

user.id is a bare text primary key (varchar(255) on MySQL, which cannot index TEXT without a prefix length) carrying no database default, because Better Auth generates the id and sends it in the insert. That is precisely what lets an app-type schema foreign-key user.id and cascade on delete. Session validation happens at two different strengths, deliberately. Next's src/proxy.ts calls getSessionCookie(request), which only asks whether the cookie is present: it runs at the edge, touches no database, and exists to bounce logged-out traffic before render. The authoritative check is auth.api.getSession({ headers }), and it runs inside the protected surface — requireUser() in src/lib/session.ts on Next, requireAuth(request) in app/lib/require-auth.ts on React Router, and the Nitro handler at server/middleware/auth.ts on Nuxt.

The last two do the real lookup on every guarded request, since neither framework has an edge proxy to peek with. The remaining emitted files are the mount: toNextJsHandler(auth) behind a [...all] route on Next, a resource route delegating to auth.handler on React Router, an h3 catch-all wrapping toWebRequest on Nuxt, plus a Vue auth client there. The auth instance imports the db client the framework already exported, so both share one pooled connection. What you inherit is ownership. Sessions join to your own tables, a user delete is a foreign-key cascade rather than a sync job, and drift arrives through your lockfile instead of a vendor's release notes.

The same ownership is the cost: password recovery only delivers if an email fragment is composed in — Better Auth's own server returns 400 "Reset password isn't enabled" until emailAndPassword.sendResetPassword is set — and rotating BETTER_AUTH_SECRET is yours to schedule.