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

# Incident protocol

> How AlgoVoi detects, communicates and recovers from chain or gateway incidents, including degraded versus down handling and the post mortem cadence.

## How we detect

The status page (Better Stack, when live) runs synthetic checks against:

* The gateway public endpoint at `api.algovoi.co.uk`
* The facilitator endpoint
* One monitor per chain's RPC or indexer (Algorand, VOI, Hedera, Stellar, Base, Solana, Tempo)

A multi-region failure on any monitor opens an incident.

## How we communicate

When a real incident is open:

* We post a short status update on the public status page within 5 minutes of detection.
* Tenants subscribed to the status page get email notifications automatically.
* For UK-business-hours incidents we update every 15 minutes until resolution.
* A post-incident summary lands within 48 hours of resolution, covering what happened, what failed, what we changed.

## What "degraded" means

A chain marked **degraded** rather than **down** means we can still verify the chain but with raised latency. Payments still settle, just slower. Don't switch off the chain in your dashboard; AlgoVoi's facilitator will route around the slow node automatically.

A chain marked **down** means the chain's primary RPC is unreachable. Payments on that chain pause. Once the RPC recovers, queued verifications drain in normal order.

## What you can do during a chain-specific incident

* **If you have multi-chain payouts configured**, your customers can pick a different chain. Recommend a working alternative in your checkout copy.
* **If you depend on one chain only**, you can either wait or temporarily pause new checkout creation on that chain in **Settings → Networks**.
* **If you have time-sensitive payments** (an agent-mediated AP2 mandate that's about to expire), the mandate's expiry can be extended on request via support.

## How to reach a human

For incidents you suspect AlgoVoi hasn't seen yet:

1. Email `support@algovoi.co.uk` with the chain, tenant ID, sample tx\_id (if any), and UTC timestamp of the first failed verification.
2. We'll reply within 1 hour during UK business hours, best-effort outside that window.
3. If the issue is multi-tenant we'll post on the status page within 5 minutes of confirming.

## Post-mortem cadence

Public post-mortems are published for any incident that:

* Caused tenant-visible payment failures for more than 30 minutes
* Touched data integrity (any payment recorded incorrectly)
* Required a security disclosure

We don't post-mortem every short outage, just the ones that mattered.

## See also

* [Support](/support) for non-incident questions
* [Trial and pricing](/trial-and-pricing) for SLA-style commitments at higher tiers