E-series Performance Analyzer Cross-site Scripting

Grafana is an open-source platform for monitoring and observability. Versions on the 8.x and 9.x branch prior to 9.0.3, 8.5.9, 8.4.10, and 8.3.10 are vulnerable to stored cross-site scripting via the Unified Alerting feature of Grafana. An attacker can exploit this vulnerability to escalate privilege from editor to admin by tricking an authenticated admin to click on a link. Versions 9.0.3, 8.5.9, 8.4.10, and 8.3.10 contain a patch. As a workaround, it is possible to disable alerting or use legacy alerting. (2022-07-15, CVE-2022-31097)

Grafana is an open-source platform for monitoring and observability. In affected versions an attacker could serve HTML content thru the Grafana datasource or plugin proxy and trick a user to visit this HTML page using a specially crafted link and execute a Cross-site Scripting (XSS) attack. The attacker could either compromise an existing datasource for a specific Grafana instance or either set up its own public service and instruct anyone to set it up in their Grafana instance. To be impacted, all of the following must be applicable. For the data source proxy: A Grafana HTTP-based datasource configured with Server as Access Mode and a URL set, the attacker has to be in control of the HTTP server serving the URL of above datasource, and a specially crafted link pointing at the attacker controlled data source must be clicked on by an authenticated user. For the plugin proxy: A Grafana HTTP-based app plugin configured and enabled with a URL set, the attacker has to be in control of the HTTP server serving the URL of above app, and a specially crafted link pointing at the attacker controlled plugin must be clocked on by an authenticated user. For the backend plugin resource: An attacker must be able to navigate an authenticated user to a compromised plugin through a crafted link. Users are advised to update to a patched version. There are no known workarounds for this vulnerability. (2022-02-08, CVE-2022-21702)

Cross site scripting is an attack where a web page executes code that is injected by an adversary. It usually appears, when users input is presented. This attack can be used to impersonate a user, take over control of the session, or even steal API keys.

The attack can be executed e.g. when you application injects the request parameter directly into the HTML code of the page returned to the user:

https://server.com/confirmation?message=Transaction+Complete

what results in:

<span>Confirmation: Transaction Complete</span>

In that case the message can be modified to become a valid Javascript code, e.g.:

https://server.com/confirmation?message=<script>dangerous javascript code here</script>

and it will be executed locally by the user's browser with full access to the user's personal application/browser data:

<span>Confirmation: <script>dangerous javascript code here</script></span>