Why Render takes time to respond after a period without visits and how to measure it

Published on 2026-11-14 by frfrand

What cold starts on Render are, how to measure the first request, and when to use a persistent execution service.

Why Render takes time to respond after a period without visits and how to measure it

Your API responds quickly during tests, but the first visit after several hours takes too long. It may be a cold start: the platform had to restart the process before serving the request. Not all slowness on Render is a cold start: DNS, SQL queries, and external services can also cause delays.

1. What happens on a free instance

According to Render's documentation, their free web services stop after 15 minutes without incoming HTTP traffic or WebSocket messages, and the first access reactivates them. Reactivation can take about a minute. Paid compute plans do not suspend due to inactivity in that way. This may change, so it is advisable to consult the current documentation before choosing a plan.

2. Measure first request and subsequent requests

Perform a controlled test: wait for an inactivity interval above the suspension threshold, make a request, and repeat immediately several times.

curl -sS -o /dev/null -w \
  'DNS=%{time_namelookup}s TCP=%{time_connect}s TLS=%{time_appconnect}s TTFB=%{time_starttransfer}s TOTAL=%{time_total}s\n' \
  https://tu-api.example/health

Record at least date, total time, and TTFB. For comparison, run the same test during a period with stable traffic. Do not confuse a request to /health with a full functional test: the endpoint may respond even if the database is down.

3. Other causes you should rule out

- Application initialization: heavy imports, migrations at startup, and model loading. - Database: initial connections, DNS, and first queries. - External services: OAuth, remote storage, and slow APIs. - Infrastructure: insufficient CPU or memory, deployments, or restarts.

To distinguish them, include traces at startup and measure queries, connections, and external calls. A fast call after restarting does not by itself prove that the problem is the plan.

4. When to pay for permanent execution?

For demos, technical tests, and applications with sporadic visits, a higher initial time may be tolerable. For webhooks, payments, real-time chat, authentication, and APIs with availability agreements, evaluate a permanently active service.

It is not recommended to rely on periodic pings designed only to prevent sleep: they can be fragile and do not solve the capacity problem.

5. Practical decision

If the delay appears only after inactivity and matches the documented behavior, compare a paid Render plan with an always-on server. If it also appears under continuous load, optimize the code and database before changing providers.

Check the ArduMaker vs. Render comparison, the FastAPI guide on Render and how to measure your API's latency.

Frequently asked questions

### Do all Render applications go to sleep? No. The documented behavior applies to free web services.

Is it the same as a slow SQL query?

No. They can produce similar symptoms; measure both paths.

Does a static website also suffer backend cold starts?

A static site served from a CDN does not have its own backend process, although the APIs it consumes may experience latency.

Source: Render, free deployments.

  • Render
  • cold start
  • latencia
  • API
  • hosting

Keep reading

All articles · VPS plans