vLLM, an inference and serving engine for large language models (LLMs), has a Regular Expression Denial of Service (ReDoS) vulnerability in the file `vllm/entrypoints/openai/tool_parsers/pythonic_tool_parser.py` of versions 0.6.4 up to but excluding 0.9.0. The root cause is the use of a highly complex and nested regular expression for tool call detection, which can be exploited by an attacker to cause severe performance degradation or make the service unavailable. The pattern contains multiple nested quantifiers, optional groups, and inner repetitions which make it vulnerable to catastrophic backtracking. Version 0.9.0 contains a patch for the issue.
vLLM is an inference and serving engine for large language models (LLMs). In version 0.8.0 up to but excluding 0.9.0, the vLLM backend used with the /v1/chat/completions OpenAPI endpoint fails to validate unexpected or malformed input in the "pattern" and "type" fields when the tools functionality is invoked. These inputs are not validated before being compiled or parsed, causing a crash of the inference worker with a single request. The worker will remain down until it is restarted. Version 0.9.0 fixes the issue.
vLLM is an inference and serving engine for large language models (LLMs). In versions 0.8.0 up to but excluding 0.9.0, hitting the /v1/completions API with a invalid json_schema as a Guided Param kills the vllm server. This vulnerability is similar GHSA-9hcf-v7m4-6m2j/CVE-2025-48943, but for regex instead of a JSON schema. Version 0.9.0 fixes the issue.
vLLM is an inference and serving engine for large language models (LLMs). Version 0.8.0 up to but excluding 0.9.0 have a Denial of Service (ReDoS) that causes the vLLM server to crash if an invalid regex was provided while using structured output. This vulnerability is similar to GHSA-6qc9-v4r8-22xg/CVE-2025-48942, but for regex instead of a JSON schema. Version 0.9.0 fixes the issue.
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.8.0 and prior to 0.8.5 are affected by a critical performance vulnerability in the input preprocessing logic of the multimodal tokenizer. The code dynamically replaces placeholder tokens (e.g., <|audio_|>, <|image_|>) with repeated tokens based on precomputed lengths. Due to inefficient list concatenation operations, the algorithm exhibits quadratic time complexity (O(n²)), allowing malicious actors to trigger resource exhaustion via specially crafted inputs. This issue has been patched in version 0.8.5.
vLLM is a high-throughput and memory-efficient inference and serving engine for LLMs. Versions starting from 0.5.2 and prior to 0.8.5 are vulnerable to denial of service and data exposure via ZeroMQ on multi-node vLLM deployment. In a multi-node vLLM deployment, vLLM uses ZeroMQ for some multi-node communication purposes. The primary vLLM host opens an XPUB ZeroMQ socket and binds it to ALL interfaces. While the socket is always opened for a multi-node deployment, it is only used when doing tensor parallelism across multiple hosts. Any client with network access to this host can connect to this XPUB socket unless its port is blocked by a firewall. Once connected, these arbitrary clients will receive all of the same data broadcasted to all of the secondary vLLM hosts. This data is internal vLLM state information that is not useful to an attacker. By potentially connecting to this socket many times and not reading data published to them, an attacker can also cause a denial of service by slowing down or potentially blocking the publisher. This issue has been patched in version 0.8.5.
Allocation of resources without limits or throttling in Windows DirectX allows an authorized attacker to deny service over a network.
An issue has been discovered in GitLab CE/EE affecting all versions from 8.15 before 18.1.5, 18.2 before 18.2.5, and 18.3 before 18.3.1 that could have could have allowed an authenticated user to cause a Denial of Service (DoS) condition by submitting URLs that generate excessively large responses.
A vulnerability in a logging API in Cisco Firepower Management Center (FMC) Software could allow an unauthenticated, remote attacker to cause the device to become unresponsive or trigger an unexpected reload. This vulnerability could also allow an attacker with valid user credentials, but not Administrator privileges, to view a system log file that they would not normally have access to. This vulnerability is due to a lack of rate-limiting of requests that are sent to a specific API that is related to an FMC log. An attacker could exploit this vulnerability by sending a high rate of HTTP requests to the API. A successful exploit could allow the attacker to cause a denial of service (DoS) condition due to the FMC CPU spiking to 100 percent utilization or to the device reloading. CPU utilization would return to normal if the attack traffic was stopped before an unexpected reload was triggered.
User-controlled operations could have allowed Denial of Service in M-Files Server before 23.4.12528.1 due to uncontrolled memory consumption.
The Image Uploader module in Liferay Portal 7.2.0 through 7.4.3.15, and older unsupported versions, and Liferay DXP 7.4 before update 16, 7.3 before update 4, 7.2 before fix pack 19, and older unsupported versions relies on a request parameter to limit the size of files that can be uploaded, which allows remote authenticated users to upload arbitrarily large files to the system's temp folder by modifying the `maxFileSize` parameter.
Bref enable serverless PHP on AWS Lambda. When Bref is used with the Event-Driven Function runtime and the handler is a `RequestHandlerInterface`, then the Lambda event is converted to a PSR7 object. During the conversion process, if the request is a MultiPart, each part is parsed and for each which contains a file, it is extracted and saved in `/tmp` with a random filename starting with `bref_upload_`. The flow mimics what plain PHP does but it does not delete the temporary files when the request has been processed. An attacker could fill the Lambda instance disk by performing multiple MultiPart requests containing files. This vulnerability is patched in 2.1.13.
The Document and Media widget In Liferay Portal 7.2.0 through 7.3.6, and older unsupported versions, and Liferay DXP 7.3 before service pack 3, 7.2 before fix pack 13, and older unsupported versions, does not limit resource consumption when generating a preview image, which allows remote authenticated users to cause a denial of service (memory consumption) via crafted PNG images.
An issue has been discovered in GitLab CE/EE affecting all versions starting from 15.11 prior to 16.9.7, starting from 16.10 prior to 16.10.5, and starting from 16.11 prior to 16.11.2. The pins endpoint is susceptible to DoS through a crafted request.
A denial of service issue was discovered in GitLab CE/EE affecting all versions starting from 13.2.4 before 15.10.8, all versions starting from 15.11 before 15.11.7, all versions starting from 16.0 before 16.0.2 which allows an attacker to cause high resource consumption using malicious test report artifacts.
OpenFGA, an authorization/permission engine, is vulnerable to a denial of service attack in versions prior to 1.4.3. In some scenarios that depend on the model and tuples used, a call to `ListObjects` may not release memory properly. So when a sufficiently high number of those calls are executed, the OpenFGA server can create an `out of memory` error and terminate. Version 1.4.3 contains a patch for this issue.
An issue has been discovered in GitLab CE/EE affecting all versions from 2.1.0 before 17.10.8, 17.11 before 17.11.4, and 18.0 before 18.0.2. A lack of input validation in HTTP responses could allow an authenticated user to cause denial of service.
Allocation of Resources Without Limits or Throttling in GitHub repository ikus060/rdiffweb prior to 2.5.5.
Mattermost versions 8.1.x <= 8.1.10, 9.6.x <= 9.6.0, 9.5.x <= 9.5.2 and 8.1.x <= 8.1.11 fail to limit the size of a request path that includes user inputs which allows an attacker to cause excessive resource consumption, possibly leading to a DoS via sending large request paths
An issue was discovered in GitLab CE/EE affecting all versions before 17.7.7, 17.8 prior to 17.8.5, and 17.9 prior to 17.9.2. where a denial of service vulnerability could allow an attacker to cause a system reboot under certain conditions.
A denial of service vulnerability in GitLab CE/EE affecting all versions from 14.1 prior to 17.6.5, 17.7 prior to 17.7.4, and 17.8 prior to 17.8.2 allows an attacker to impact the availability of GitLab via unbounded symbol creation via the scopes parameter in a Personal Access Token.
Allocation of Resources Without Limits or Throttling vulnerability in Kron Technologies Kron PAM allows HTTP DoS.This issue affects Kron PAM: before 3.7.
ReportPortal is an AI-powered test automation platform. Prior to version 5.10.0 of the `com.epam.reportportal:service-api` module, corresponding to ReportPortal version 23.2, the ReportPortal database becomes unstable and reporting almost fully stops except for small launches with approximately 1 test inside when the test_item.path field is exceeded the allowable `ltree` field type indexing limit (path length>=120, approximately recursive nesting of the nested steps). REINDEX INDEX path_gist_idx and path_idx aren't helped. The problem was fixed in `com.epam.reportportal:service-api` module version 5.10.0 (product release 23.2), where the maximum number of nested elements were programmatically limited. A workaround is available. After deletion of the data with long paths, and reindexing both indexes (path_gist_idx and path_idx), the database becomes stable and ReportPortal works properly.
An issue has been discovered in GitLab EE affecting all versions from 13.3.0 prior to 16.6.7, 16.7 prior to 16.7.5, and 16.8 prior to 16.8.2 which allows an attacker to do a resource exhaustion using GraphQL `vulnerabilitiesCountByDay`
A vulnerable API method in M-Files Server before 23.12.13195.0 allows for uncontrolled resource consumption. Authenticated attacker can exhaust server storage space to a point where the server can no longer serve requests.
A denial of service vulnerability was identified in GitLab CE/EE, affecting all versions from 15.11 prior to 16.6.7, 16.7 prior to 16.7.5 and 16.8 prior to 16.8.2 which allows an attacker to spike the GitLab instance resource usage resulting in service degradation.
A flaw was found in CRI-O that involves an experimental annotation leading to a container being unconfined. This may allow a pod to specify and get any amount of memory/cpu, circumventing the kubernetes scheduler and potentially resulting in a denial of service in the node.
api/account/register in the TH Wildau COVID-19 Contact Tracing application through 2021-09-01 has Incorrect Access Control. An attacker can interfere with tracing of infection chains by creating 500 random users within 2500 seconds.
A lack of rate limiting in the 'Forgot Password' feature of PHPJabbers Event Ticketing System v1.0 allows attackers to send an excessive amount of email for a legitimate user, leading to a possible Denial of Service (DoS) via a large amount of generated e-mail messages.
An issue has been discovered in GitLab EE affecting all versions starting from 10.5 before 16.4.3, all versions starting from 16.5 before 16.5.3, all versions starting from 16.6 before 16.6.1. It was possible for an attacker to cause a client-side denial of service using malicious crafted mermaid diagram input.
IBM Db2 for Linux, UNIX and Windows (includes Db2 Connect Server) 10.5, 11.1, and 11.5 could allow an authenticated user with CONNECT privileges to cause a denial of service using a specially crafted query. IBM X-Force ID: 272644.
A flaw was found in glusterfs server through versions 4.1.4 and 3.1.2 which allowed repeated usage of GF_META_LOCK_KEY xattr. A remote, authenticated attacker could use this flaw to create multiple locks for single inode by using setxattr repetitively resulting in memory exhaustion of glusterfs server node.
Jmix is a set of libraries and tools to speed up Spring Boot data-centric application development. In versions 1.0.0 to 1.6.1 and 2.0.0 to 2.3.4, the local file storage implementation does not restrict the size of uploaded files. An attacker could exploit this by uploading excessively large files, potentially causing the server to run out of space and return HTTP 500 error, resulting in a denial of service. This issue has been patched in versions 1.6.2 and 2.4.0. A workaround is provided on the Jmix documentation website.
An issue has been discovered in GitLab CE/EE affecting all versions from 10.7 before 17.11.5, 18.0 before 18.0.3, and 18.1 before 18.1.1 that could have allowed authenticated attackers to create a DoS condition by sending crafted GraphQL requests.
CUBA Platform is a high level framework for enterprise applications development. Prior to version 7.2.23, the local file storage implementation does not restrict the size of uploaded files. An attacker could exploit this by uploading excessively large files, potentially causing the server to run out of space and return HTTP 500 error, resulting in a denial of service. This issue has been patched in version 7.2.23. A workaround is provided on the Jmix documentation website.
Mattermost versions 10.4.x <= 10.4.2, 10.5.x <= 10.5.0, 9.11.x <= 9.11.10 fail to validate the uniqueness and quantity of task actions within the UpdateRunTaskActions GraphQL operation, which allows an attacker to create task items containing an excessive number of actions triggered by specific posts, overloading the server and leading to a denial-of-service (DoS) condition.
When parsing a multipart form (either explicitly with Request.ParseMultipartForm or implicitly with Request.FormValue, Request.PostFormValue, or Request.FormFile), limits on the total size of the parsed form were not applied to the memory consumed while reading a single form line. This permits a maliciously crafted input containing very long lines to cause allocation of arbitrarily large amounts of memory, potentially leading to memory exhaustion. With fix, the ParseMultipartForm function now correctly limits the maximum size of form lines.
It was found in Ceph versions before 13.2.4 that authenticated ceph RGW users can cause a denial of service against OMAPs holding bucket indices.
XGrammar is an open-source library for efficient, flexible, and portable structured generation. Prior to 0.1.18, Xgrammar includes a cache for compiled grammars to increase performance with repeated use of the same grammar. This cache is held in memory. Since the cache is unbounded, a system making use of xgrammar can be abused to fill up a host's memory and case a denial of service. For example, sending many small requests to an LLM inference server with unique JSON schemas would eventually cause this denial of service to occur. This vulnerability is fixed in 0.1.18.
An authenticated malicious user could initiate multiple concurrent requests, each requesting multiple dashboard exports, leading to a possible denial of service. This issue affects Apache Superset: before 3.0.0
matrix-media-repo is an open-source multi-domain media repository for Matrix. Versions 1.2.6 and earlier of matrix-media-repo do not properly handle malicious images which are crafted to be small in file size, but large in complexity. A malicious user could upload a relatively small image in terms of file size, using particular image formats, which expands to have extremely large dimensions during the process of thumbnailing. The server can be exhausted of memory in the process of trying to load the whole image into memory for thumbnailing, leading to denial of service. Version 1.2.7 has a fix for the vulnerability.
An issue has been discovered in GitLab CE/EE affecting all versions from 10.2 before 17.10.7, 17.11 before 17.11.3, and 18.0 before 18.0.1. A lack of input validation in the Kubernetes integration could allow an authenticated user to cause denial of service..
Discourse is an open-source discussion platform. Prior to version 3.1.1 of the `stable` branch and version 3.2.0.beta1 of the `beta` and `tests-passed` branches, a malicious admin could create extremely large icons sprites, which would then be cached in each server process. This may cause server processes to be killed and lead to downtime. The issue is patched in version 3.1.1 of the `stable` branch and version 3.2.0.beta1 of the `beta` and `tests-passed` branches. This is only a concern for multisite installations. No action is required when the admins are trusted.
Discourse is an open-source discussion platform. Prior to version 3.1.1 of the `stable` branch and version 3.2.0.beta1 of the `beta` and `tests-passed` branches, importing a remote theme loads their assets into memory without enforcing limits for file size or number of files. The issue is patched in version 3.1.1 of the `stable` branch and version 3.2.0.beta1 of the `beta` and `tests-passed` branches. There are no known workarounds.
IBM Db2 for Linux, UNIX and Windows (includes DB2 Connect Server) 11.5.0 through 11.5.9 and 12.1.0 through 12.1.1 could allow an authenticated user to cause a denial of service when using Q replication due to the improper allocation of CPU resources.
An issue has been discovered in GitLab CE/EE affecting all versions before 17.10.7, 17.11 before 17.11.3, and 18.0 before 18.0.1. A lack of proper validation in GitLab could allow an authenticated user to cause a denial of service condition.
An issue was discovered in Datalust Seq before 2024.3.13545. Expansion of identifiers in message templates can be used to bypass the system "Event body limit bytes" setting, leading to increased resource consumption. With sufficiently large events, there can be disk space exhaustion (if saved to disk) or a termination of the server process with an out-of-memory error.
An issue has been discovered in GitLab CE/EE affecting all versions from 11.6 before 18.0.6, 18.1 before 18.1.4, and 18.2 before 18.2.2 that could have allowed an authenticated user to cause a denial of service condition by creating specially crafted content that consumes excessive server resources when processed.
A vulnerability in the web application of ctrlX OS allows a remote authenticated (low-privileged) attacker to induce a Denial-of-Service (DoS) condition on the device via multiple crafted HTTP requests. In the worst case, a full power cycle is needed to regain control of the device.
In Concrete CMS (formerly concrete5) below 8.5.10 and between 9.0.0 and 9.1.2, the authTypeConcreteCookieMap table can be filled up causing a denial of service (high load).