Rate Limits
Request limits can depend on the LCE deployment, gateway, environment and integration agreement. Build clients that control their request volume even when a fixed public quota is not published.
Current contract
Reduce unnecessary traffic
Paginate
Use moderate page sizes and process one page before requesting the next.
Filter
Query only the IDs, statuses, cultures and date ranges required by the workflow.
Cache
Cache stable configuration such as language metadata according to the project’s freshness requirements.
Batch responsibly
Use bulk operations only where documented and split large jobs into controlled chunks.
Cap concurrency
Set a maximum number of in-flight requests per service and environment.
Observe
Measure latency, errors, retries and request volume before increasing throughput.
If the gateway returns 429
HTTP/1.1 429 Too Many Requests
Retry-After: 10
- Pause new requests for the affected scope.
- Honor
Retry-Afterwhen the response provides it. - Otherwise use capped exponential backoff with jitter.
- Retry only operations that are safe to repeat.
- Reduce concurrency if throttling continues.
The response above is an integration pattern, not a guarantee that every LCE deployment returns the same status or header.
Capped backoff example
function backoffMs(attempt, retryAfterSeconds) {
if (retryAfterSeconds != null) {
return retryAfterSeconds * 1000;
}
const base = Math.min(1000 * 2 ** attempt, 30000);
const jitter = Math.floor(Math.random() * 500);
return base + jitter;
}
Cap the number of attempts and total retry duration. Do not retry validation failures, authorization failures or declined business outcomes.
Plan high-volume integrations
| Question | Why it matters |
|---|---|
| Which environment and gateway will receive traffic? | Limits and infrastructure capacity can differ between development and production. |
| What is the expected peak and sustained volume? | Average traffic alone does not reveal burst behavior. |
| Which operations have side effects? | Retries require a stricter policy for creates, processes and provider calls. |
| Can requests be queued or scheduled? | Background workloads can be smoothed instead of competing with interactive traffic. |
| What monitoring is available? | Latency and throttling signals are required for safe capacity tuning. |