We moved AI OGO to a new server and restored data from a backup. Here's what happened and how we're preparing for future outages.
AI OGO is up and running again. We're sorry for the long outage, and thank you for waiting for us. Here's how we brought the service back and what we're doing so we can recover faster next time.
What happened?
On the night of November 1–2, our hosting provider Beget suffered an outage. According to the provider, it was caused by a cyberattack on its infrastructure. While the old server remained inaccessible, we deployed AI OGO on a new one. See Beget's incident timeline.
How did we restore the service?
Our backup was stored separately from the main server. That meant we could restore the database from the latest available copy, including account data and saved history, and bring the site back online while the old server was still down.
Why are we preparing for future outages?
Companies around the world are facing cyberattacks, and AI is helping attackers move faster. It makes finding vulnerabilities easier and lets them produce convincing phishing messages at scale. The UK's National Cyber Security Centre expects these threats to grow more frequent and more severe as AI advances. We may be seeing only the beginning of this trend. Read the NCSC assessment of AI's impact on cyber threats.
How will our Continuity Plan work?
We're setting up a backup site with a different provider in a different location. If one site goes down because of an attack or a technical failure, we can bring the service up on the other. To make this possible, we're configuring regular backup updates and preparing to test the failover in advance.
If a serious outage happens, our goal is to get AI OGO back up within a few hours. This time, having a separate backup already helped us recover. The next step is to rehearse the failover so we can act immediately if something goes wrong.
Thank you for staying with us. Open AI OGO.
