Skip to content

Un-sanitized metric name or labels can be used to take over exported metrics

Moderate severity GitHub Reviewed Published Mar 29, 2024 in swift-server/swift-prometheus

Package

swift github.com/swift-server/swift-prometheus (Swift)

Affected versions

= 2.0.0-alpha.1

Patched versions

2.0.0-alpha.2

Description

Impact

In code which applies un-sanitized string values into metric names or labels, like this:

let lang = try? request.query-get(String.self, at: "lang")
Counter (
  label: "language", 
  dimensions: [("lang", lang ?? "unknown" )]
)

an attacker could make use of this and send a ?lang query parameter containing newlines, } or similar characters which can lead to the attacker taking over the exported format -- including creating unbounded numbers of stored metrics, inflating server memory usage, or causing "bogus" metrics.

Patches

The default strategy to sanitize labels was moved deeper into the library, preventing illegal characters from appearing in name, label keys and values.

Metric names and label names are now validated against the following requirement: [a-zA-Z_:][a-zA-Z0-9_:]* (for metric names) and [a-zA-Z_][a-zA-Z0-9_]* (for metric label names). Label values are not validated as they are allowed to contain any unicode characters. Developers must validate labels themselves and not allow malicious input.

The approach taken here mirrors the approach taken in the Go reference implementation.

Discussion

It is strongly discouraged to use un-sanitized user input as names or labels in general, because they can lead to un-bounded growth of metrics, even as this vulnerability is patched and result in a Denial-of-Service attack opportunity -- regardless how well the library is sanitizing the inputs. We strongly recommend only using a sanitized set of values for your metrics names and labels. E.g., a "lang" label, should only use an expected set of values that can be used, and ignore other ones -- otherwise a determined attacker could create one metric per different label key, leading to unbounded memory use growth as metrics with distinct values must be kept in memory.

Validating label values:

The library will NOT automatically validate and replace strings offered as label values.
Developers must validate label values themselves, and it is strongly recommended to only accept a well known set of values.

It is possible to configure the PrometheusSanitizer to apply whatever validation you deem necessary:

let mySanitizer = PrometheusSanitizer { metricName, labels in
  // ... your logic here ...
  (metricName, labels)
}

let registry = PrometheusCollectorRegistry(sanitizer: mySanitizer)
let factory = PrometheusMetricsFactory(factory: registry)

// swift-metrics
MetricsSystem.bootstrap(factory)

Workarounds

Developers must validate user input before using it as metric names, label names or values. This follows common practice of not trusting any user input without sanitization.

Credits

We would like to thank Jonas Dörr for bringing out attention to the issue.

References

@ktoso ktoso published to swift-server/swift-prometheus Mar 29, 2024
Published by the National Vulnerability Database Mar 29, 2024
Published to the GitHub Advisory Database Mar 29, 2024
Reviewed Mar 29, 2024

Severity

Moderate

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Network
Attack complexity
High
Privileges required
None
User interaction
None
Scope
Unchanged
Confidentiality
None
Integrity
High
Availability
None

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:H/A:N

EPSS score

0.043%
(10th percentile)

Weaknesses

CVE ID

CVE-2024-28867

GHSA ID

GHSA-x768-cvr2-345r
Loading Checking history
See something to contribute? Suggest improvements for this vulnerability.