Who is affected
Metabase instances are vulnerable to unauthenticated SQL injection that grants administrator access to the application itself. This affects any organisation running Metabase, whether as a self-hosted deployment or cloud service, that has not applied vendor mitigations. The record does not specify which versions are affected or whether a patch is available.
How to check whether this touches you
- Inventory all Metabase instances in your estate, including development, staging and production deployments.
- For each instance, determine whether it is reachable from the public internet or from untrusted networks; check your firewall rules and reverse proxy configurations.
- Check the Metabase admin panel (Settings > About) or query the running container/process to confirm the installed version; note that version numbers alone do not prove whether a fix has been backported.
- Test whether unauthenticated access to the application is possible; if you can reach the login page without credentials, the network exposure exists.
What to do
- Contact Metabase immediately to obtain patch or mitigation guidance for your specific version; do not assume a public advisory exists yet.
- If patching is not immediately available, restrict network access to Metabase to trusted IP ranges and authenticated VPN only; place the instance behind a WAF or reverse proxy that requires authentication.
- Enable and retain all application logs, database query logs, and network access logs for at least 90 days; configure alerting on failed and successful login attempts, especially to the admin account.
- If your Metabase instance is internet-facing and you cannot patch or restrict access within 72 hours, escalate to your change advisory board and consider taking the instance offline until mitigations are in place.
- Review the credentials Metabase holds for connected databases; plan to rotate them after patching is confirmed.
If you find you were exposed
Exploitation of SQL injection typically occurs before public disclosure, so assume your logs may contain evidence of attack attempts weeks or months in the past. Check authentication logs, database query logs, and any audit records for anomalous queries, failed login storms, or admin account creation during periods before today. If log retention is less than 90 days, you may be unable to establish the start of exposure; document this gap and plan longer retention for future incidents.