> ## Documentation Index
> Fetch the complete documentation index at: https://docs.generect.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Reliability

> Uptime, status page, and how the API behaves under failure

## Live status & uptime

Current availability, **90-day uptime history**, and incident post-mortems are public at [status.generect.com](https://status.generect.com) — subscribe there for incident notifications.

Contractual SLAs are available on Enterprise agreements — [contact sales](https://calendly.com/andrii-kolpakov/_sales_api_insight).

## What to expect from each mode

* **Database (cached)** operations answer from Generect's own infrastructure — sub-second, independent of live sources.
* **Real-time** operations depend on live upstream sources: latency is seconds to a minute (see [recommended timeouts](/api-reference/limits)), and under exceptional upstream load they can return `429` — back off and retry.

## Failure semantics

* Errors use one JSON envelope with predictable status codes — see [Errors](/api-reference/errors).
* `429` and `5xx` are transient: retry with exponential backoff. Other `4xx` mean the request must change.
* Credits are **reserved when a request or bulk job is accepted** — a failed request is refunded automatically, and an accepted bulk job always runs to completion even if your balance later hits \$0 (see [Pricing](/billing/pricing)).
* Bulk jobs report `error` status instead of hanging; results are retained for 24 hours.