Tolgee - too many requests

[durino13]

Hi everyone,

we’re running a self-hosted Tolgee instance in production (Docker Compose) and occasionally the Tolgee UI shows: “Too many requests. Try again later.” We also use Tolgee’s REST API from our app to fetch translations.

We tried to change the configuration to:

tolgee:
rate-limits:
global-limits: false
user-request-limit: 100000
endpoint-limits: false

Now we will wait, if the problem occures again ..

Context

Tolgee is behind a public domain (e.g. https://tolgee.example.com).
Our frontend (Nuxt SSR) and backend application fetches translations from Tolgee (/v2/projects//translations/?filterTag=) and caches them in Redis with a 24h TTL.
We also have some backend endpoints which call a translation helper that reads slugs/translations from Redis; on cache miss it triggers a refresh from Tolgee (fetch-all translations for a tag and store them in Redis).

Questions

Does Tolgee provide any built-in way (UI or API) to see request statistics (per endpoint / per IP), or is this only possible via reverse-proxy access logs?
Any tips on how to conclusively verify whether 'too many requests' is generated by Tolgee vs. a reverse proxy?
Any other suggestions, how to debug this issue?

Thanks in advance for your insights!

[Anty0]

Hey @durino13,

Great questions! Let me go through them one by one.

1. Request Statistics via Prometheus

Tolgee exposes a Prometheus-compatible metrics endpoint at /actuator/prometheus. You can point your Grafana/Prometheus setup to scrape this directly — it gives you standard HTTP metrics (request counts, response times, status codes per endpoint) via Micrometer, plus some Tolgee-specific metrics.

Just add http://<tolgee-host>:<port>/actuator/prometheus as a scrape target in your Prometheus config.

2. Tolgee 429 vs Reverse Proxy 429

Tolgee's rate limit response has a specific JSON body format:

{
  "message": "...",
  "retryAfter": 12345,
  "global": true
}

• retryAfter — milliseconds until you can retry
• global — true for global IP/user limits, false for per-endpoint limits

If the 429 response body doesn't match this format, it's coming from your reverse proxy.

3. Debugging Tips

You can enable DEBUG logging for the rate limit classes to see exactly which limits are being triggered:

logging:
  level:
    io.tolgee.security.ratelimit: DEBUG

4. Your Config

Your config looks correct for the /v2/projects/<id>/translations/<lang> endpoint. This endpoint is only subject to global rate limits (IP-based and user-based) — it doesn't have any additional per-endpoint limits. Setting global-limits: false disables both of those, so the 429s should stop.

The user-request-limit: 100000 and endpoint-limits: false in your config are technically redundant (since global-limits: false already disables the limits affecting this endpoint), but they don't cause any harm.