Headline
CVE-2022-24812: Build software better, together
Grafana is an open-source platform for monitoring and observability. When fine-grained access control is enabled and a client uses Grafana API Key to make requests, the permissions for that API Key are cached for 30 seconds for the given organization. Because of the way the cache ID is constructed, the consequent requests with any API Key evaluate to the same permissions as the previous requests. This can lead to an escalation of privileges, when for example a first request is made with Admin permissions, and the second request with different API Key is made with Viewer permissions, the second request will get the cached permissions from the previous Admin, essentially accessing higher privilege than it should. The vulnerability is only impacting Grafana Enterprise when the fine-grained access control beta feature is enabled and there are more than one API Keys in one organization with different roles assigned. All installations after Grafana Enterprise v8.1.0-beta1 should be upgraded as soon as possible. As an alternative, disable fine-grained access control will mitigate the vulnerability.
Today we are releasing Grafana Enterprise 8.3.6. This patch release includes HIGH severity security fix for Grafana Enterprise fine-grained access control API Key privilege escalation.
Release v.8.4.6, only containing security fixes:
- Download Grafana Enterprise 8.4.6
- Release notes
Privilige Escalation
On 3rd of April, during an internal security audit, we discovered a security vulnerability which impacts Grafana Enterprise instances which have fine-grained access control beta feature enabled.
We believe that this vulnerability is rated at CVSS 8.0 (CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:C/C:H/I:H/A:H).
Pre-conditions
- Running Grafana Enterprise 8.1.0-beta1 - 8.4.5
- Fine-grained access control is enabled
- There are at least 2 API Keys with different roles (e.g. one with Admin, one with Editor).
Summary
When fine-grained access control is enabled and a client uses Grafana API Key to make requests, the permissions for that API Key are cached for 30 seconds for the given organization. Because of the way the cache ID is constructed, the consequent requests with any API Key evaluate to the same permissions as the previous requests. This can lead to an escalation of privileges, when for example a first request is made with Admin permissions, and the second request with different API Key is made with Viewer permissions, the second request will get the cached permissions from the previous Admin, essentially accessing higher privilege than it should.
Impact
The vulnerability is only impacting Grafana Enterprise when the fine-grained access control beta feature is enabled and there are more than one API Keys in one organization with different roles assigned.
Affected versions with HIGH severity
All Grafana Enterprise instances with 8.1.0-beta1 - 8.4.5 versions are affected by this vulnerability.
Solutions and mitigations
All installations after Grafana Enterprise v8.1.0-beta1 should be upgraded as soon as possible. As an alternative, disabling fine-grained access control will mitigate the vulnerability.
Appropriate patches have been applied to Grafana Cloud and as always, we closely coordinated with all cloud providers licensed to offer Grafana Pro. They have received early notification under embargo and confirmed that their offerings are secure at the time of this announcement. This is applicable to Amazon Managed Grafana and Azure’s Grafana as a service offering.
Timeline and postmortem
Here is a detailed timeline starting from when we originally learned of the issue. All times in UTC.
- 2022-04-03 13:02 A potential Issue related to fine-grained access control escalated internally
- 2022-04-03 12:07 Issue escalated and the vulnerability confirmed reproducible
- 2022-04-03 14:30 Decision is made to release a private patch
- 2022-04-04 09:06 CVE requested
- 2022-04-04 12:40 Private release planned for 2022-04-05, and public release planned for 2022-04-12
- 2022-04-04 12:40 GitHub has issued CVE-2022-24812
- 2022-04-05 12:00 Private release
- 2022-04-12 12:00 Public release
Reporting security issues
If you think you have found a security vulnerability, please send a report to [email protected]. This address can be used for all of Grafana Labs’s open source and commercial products (including but not limited to Grafana, Grafana Cloud, Grafana Enterprise, and grafana.com). We can accept only vulnerability reports at this address.
Please encrypt your message to us; please use our PGP key. The key fingerprint is:
F988 7BEA 027A 049F AE8E 5CAA D125 8932 BE24 C5CA
The key is available from keyserver.ubuntu.com.
Grafana Labs will send you a response indicating the next steps in handling your report. After the initial reply to your report, the security team will keep you informed of the progress towards a fix and full announcement, and may ask for additional information or guidance.
Important: We ask you to not disclose the vulnerability before it have been fixed and announced, unless you received a response from the Grafana Labs security team that you can do so.
Security announcements
We maintain a category on the community site called Security Announcements,
where we will post a summary, remediation, and mitigation details for any patch containing security fixes.
You can also subscribe to email updates to this category if you have a grafana.com account and sign on to the community site or track updates via an RSS feed.
Related news
An issue in the component Item_subselect::init_expr_cache_tracker of MariaDB Server v10.6 and below was discovered to allow attackers to cause a Denial of Service (DoS) via specially crafted SQL statements.
MariaDB Server v10.6 and below was discovered to contain an use-after-free in the component my_strcasecmp_8bit, which is exploited via specially crafted SQL statements.
An issue in the component Create_tmp_table::finalize of MariaDB Server v10.7 and below was discovered to allow attackers to cause a Denial of Service (DoS) via specially crafted SQL statements.
An issue in the component Field::set_default of MariaDB Server v10.6 and below was discovered to allow attackers to cause a Denial of Service (DoS) via specially crafted SQL statements.
An issue in the component Arg_comparator::compare_real_fixed of MariaDB Server v10.6.2 and below was discovered to allow attackers to cause a Denial of Service (DoS) via specially crafted SQL statements.
An issue in the component Used_tables_and_const_cache::used_tables_and_const_cache_join of MariaDB Server v10.7 and below was discovered to allow attackers to cause a Denial of Service (DoS) via specially crafted SQL statements.
MariaDB Server v10.7 and below was discovered to contain a global buffer overflow in the component decimal_bin_size, which is exploited via specially crafted SQL statements.
MariaDB Server v10.6.3 and below was discovered to contain an use-after-free in the component Item_func_in::cleanup(), which is exploited via specially crafted SQL statements.
Subversion's mod_dav_svn is vulnerable to memory corruption. While looking up path-based authorization rules, mod_dav_svn servers may attempt to use memory which has already been freed. Affected Subversion mod_dav_svn servers 1.10.0 through 1.14.1 (inclusive). Servers that do not use mod_dav_svn are not affected.
MinIO is a High Performance Object Storage released under GNU Affero General Public License v3.0. A security issue was found where an non-admin user is able to create service accounts for root or other admin users and then is able to assume their access policies via the generated credentials. This in turn allows the user to escalate privilege to that of the root user. This vulnerability has been resolved in pull request #14729 and is included in `RELEASE.2022-04-12T06-55-35Z`. Users unable to upgrade may workaround this issue by explicitly adding a `admin:CreateServiceAccount` deny policy, however, this, in turn, denies the user the ability to create their own service accounts as well.
An arbitrary file upload vulnerability in the file upload module of PayloadCMS v0.15.0 allows attackers to execute arbitrary code via a crafted SVG file.
An arbitrary file upload vulnerability in the file upload module of Ghost CMS v4.42.0 allows attackers to execute arbitrary code via a crafted file.
An arbitrary file upload vulnerability in the file upload module of Strapi v4.1.5 allows attackers to execute arbitrary code via a crafted file.
A SQL injection issue was discovered in QuerySet.explain() in Django 2.2 before 2.2.28, 3.2 before 3.2.13, and 4.0 before 4.0.4. This occurs by passing a crafted dictionary (with dictionary expansion) as the **options argument, and placing the injection payload in an option name.
Nokogiri is an open source XML and HTML library for Ruby. Nokogiri `< v1.13.4` contains an inefficient regular expression that is susceptible to excessive backtracking when attempting to detect encoding in HTML documents. Users are advised to upgrade to Nokogiri `>= 1.13.4`. There are no known workarounds for this issue.
GoCD is an open source a continuous delivery server. The bundled gocd-ldap-authentication-plugin included with the GoCD Server fails to correctly escape special characters when using the username to construct LDAP queries. While this does not directly allow arbitrary LDAP data exfiltration, it can allow an existing LDAP-authenticated GoCD user with malicious intent to construct and execute malicious queries, allowing them to deduce facts about other users or entries within the LDAP database (e.g alternate fields, usernames, hashed passwords etc) through brute force mechanisms. This only affects users who have a working LDAP authorization configuration enabled on their GoCD server, and only is exploitable by users authenticating using such an LDAP configuration. This issue has been fixed in GoCD 22.1.0, which is bundled with gocd-ldap-authentication-plugin v2.2.0-144.
PrivateBin is minimalist, open source online pastebin clone where the server has zero knowledge of pasted data. In PrivateBin < v1.4.0 a cross-site scripting (XSS) vulnerability was found. The vulnerability is present in all versions from v0.21 of the project, which was at the time still called ZeroBin. The issue is caused by the fact that SVGs can contain JavaScript. This can allow an attacker to execute code, if the user opens a paste with a specifically crafted SVG attachment, and interacts with the preview image and the instance isn't protected by an appropriate content security policy. Users are advised to either upgrade to version 1.4.0 or to ensure the content security policy of their instance is set correctly.
Elide is a Java library that lets you stand up a GraphQL/JSON-API web service with minimal effort. When leveraging the following together: Elide Aggregation Data Store for Analytic Queries, Parameterized Columns (A column that requires a client provided parameter), and a parameterized column of type TEXT. There is the potential for a hacker to provide a carefully crafted query that would bypass server side authorization filters through SQL injection. A recent patch to Elide 6.1.2 allowed the '-' character to be included in parameterized TEXT columns. This character can be interpreted as SQL comments ('--') and allow the attacker to remove the WHERE clause from the generated query and bypass authorization filters. A fix is provided in Elide 6.1.4. The vulnerability only exists for parameterized columns of type TEXT and only for analytic queries (CRUD is not impacted). Workarounds include leveraging a different type of parameterized column (TIME, MONEY, etc) or not leveraging parameteriz...
JHipster is a development platform to quickly generate, develop, & deploy modern web applications & microservice architectures. SQL Injection vulnerability in entities for applications generated with the option "reactive with Spring WebFlux" enabled and an SQL database using r2dbc. Applications created without "reactive with Spring WebFlux" and applications with NoSQL databases are not affected. Users who have generated a microservice Gateway using the affected version may be impacted as Gateways are reactive by default. Currently, SQL injection is possible in the findAllBy(Pageable pageable, Criteria criteria) method of an entity repository class generated in these applications as the where clause using Criteria for queries are not sanitized and user input is passed on as it is by the criteria. This issue has been patched in v7.8.1. Users unable to upgrade should be careful when combining criterias and conditions as the root of the issue lies in the `EntityManager.java` class when cre...
Garden is an automation platform for Kubernetes development and testing. In versions prior to 0.12.39 multiple endpoints did not require authentication. In some operating modes this allows for an attacker to gain access to the application erroneously. The configuration is leaked through the /api endpoint on the local server that is responsible for serving the Garden dashboard. At the moment, this server is accessible to 0.0.0.0 which makes it accessible to anyone on the same network (or anyone on the internet if they are on a public, static IP). This may lead to the ability to compromise credentials, secrets or environment variables. Users are advised to upgrade to version 0.12.39 as soon as possible. Users unable to upgrade should use a firewall blocking access to port 9777 from all untrusted network machines.