Observability

Logs + errors + traces, all correlated by request-id. Free tiers cover us; upgrade only when scale forces.

Recommended stack

Logging — pino

pnpm --filter @spacehub/api add pino pino-pretty pino-http
pnpm --filter @spacehub/workers add pino

// apps/api/src/lib/logger.ts
import pino from "pino";
export const logger = pino({
  level: process.env.LOG_LEVEL ?? "info",
  transport: process.env.NODE_ENV === "development"
    ? { target: "pino-pretty", options: { colorize: true } }
    : undefined,                                              // JSON in prod
  base: { service: "api", env: process.env.NODE_ENV },
});

Per-request child logger with request-id:

// apps/api/src/middleware/request-logger.ts
import { createMiddleware } from "hono/factory";
import { AsyncLocalStorage } from "node:async_hooks";
import { logger } from "../lib/logger";

export const reqLogStore = new AsyncLocalStorage<pino.Logger>();

export const requestLogger = createMiddleware(async (c, next) => {
  const reqId = c.get("requestId");
  const child = logger.child({ reqId, method: c.req.method, path: c.req.path });
  const start = Date.now();
  await reqLogStore.run(child, next);
  child.info({ ms: Date.now() - start, status: c.res.status }, "request");
});

// anywhere in handler stack: const log = reqLogStore.getStore() ?? logger;

Error tracking — Sentry

@sentry/nextjs v9+ supports Next 15 App Router + Turbopack + onRequestError. @sentry/hono beta in 2026 replaces deprecated community @hono/sentry.

pnpm --filter @spacehub/api add @sentry/node @sentry/hono
pnpm --filter @spacehub/web add @sentry/nextjs

// apps/api/src/lib/sentry.ts
import * as Sentry from "@sentry/node";
import { sentry } from "@sentry/hono";
Sentry.init({ dsn: process.env.SENTRY_DSN, tracesSampleRate: 0.1, profilesSampleRate: 0.1 });

app.use("*", sentry({ dsn: process.env.SENTRY_DSN! }));

Self-host fallback: GlitchTip (Sentry-protocol compatible) on Coolify if you want everything on one VPS.

Tracing — OpenTelemetry

pnpm --filter @spacehub/api add @opentelemetry/auto-instrumentations-node @hono/otel

// apps/api/src/lib/telemetry.ts (must import BEFORE Hono)
import { NodeSDK } from "@opentelemetry/sdk-node";
import { getNodeAutoInstrumentations } from "@opentelemetry/auto-instrumentations-node";
import { OTLPTraceExporter } from "@opentelemetry/exporter-trace-otlp-http";

new NodeSDK({
  traceExporter: new OTLPTraceExporter({ url: process.env.OTEL_EXPORTER_OTLP_ENDPOINT }),
  instrumentations: [getNodeAutoInstrumentations({
    "@opentelemetry/instrumentation-fs": { enabled: false },     // noisy
  })],
}).start();

// In Hono:
import { otel } from "@hono/otel";
app.use("*", otel());

Auto-instrumentations cover: http, pg (Drizzle queries via the driver), ioredis, undici (fetch). Drizzle has no first-party OTel integration — wrap db.execute in a thin span helper for query-level naming if needed.

Metrics

OTel auto-instrumentations export to OTLP. Grafana Cloud accepts OTLP natively. Custom metrics via @opentelemetry/api:

import { metrics } from "@opentelemetry/api";
const meter = metrics.getMeter("spacehub");
const billsGenerated = meter.createCounter("bills_generated_total");
billsGenerated.add(1, { ownerId, period });

Uptime + heartbeats

• BetterStack monitors /v2/health from multiple regions (incl. Tokyo for Mongolia-relevance).
• Scheduler heartbeat: BullMQ JobScheduler emits ping every 5 min; /healthz/scheduler returns last-emit timestamp. BetterStack alerts if gap > 10 min.

Feature flags + analytics — PostHog

One tool for flags + product analytics + (optional) session replay. Self-host on Coolify if event volume crosses 1M/mo free cap.

pnpm --filter @spacehub/web add posthog-js
pnpm --filter @spacehub/api add posthog-node

Build steps

1. Phase 0: pino + request-id middleware (already in scaffold).
2. Sentry init in API + web (Phase 1).
3. OTel SDK + auto-instrumentations + Grafana Cloud free tier (Phase 1).
4. BetterStack health monitors + scheduler heartbeat (Phase 1).
5. PostHog (Phase 5 or whenever first owner-facing feature flag is needed).

Open questions

• Trace sample rate: 10% default (cost-friendly). Bump to 100% for the first month to baseline.
• PII scrubbing in logs: redact email, phone, TIN by default. Use pino's redact option.
• Log retention: 30 days at the Loki free tier is enough for ops; longer = paid.
• Audit log (separate from operational logs): durable Postgres table for compliance — auth events, money mutations, role changes. Recommended yes.

Sources

• Sentry for Hono
• Sentry for Next.js
• @hono/otel