EUVD-2026-41077
GHSA-w52p-h72q-gcpf
Severity by source
AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
Network-reachable, low-complexity memory-exhaustion DoS with availability-only impact (A:H, C/I:N); PR:N follows the vendor vector though upload endpoints may in practice require an agent API key.
Primary rating from NVD.
CVSS VectorNVD
Lifecycle Timeline
6DescriptionNVD
Allocation of Resources Without Limits or Throttling (CWE-770) in Fleet Server can lead to a denial of service via Excessive Allocation (CAPEC-130). An attacker can submit a specially crafted request to an upload endpoint that causes excessive memory consumption, which may render Fleet Server unavailable.
AnalysisAI
Denial of service in Elastic Fleet Server (versions 8.0.0-8.19.10 and 9.0.0-9.2.4) allows remote attackers to exhaust memory by sending a specially crafted request to an upload endpoint, driving excessive allocation until the server becomes unresponsive. The flaw is a resource-throttling failure (CWE-770) with availability-only impact and no confidentiality or integrity loss. No public exploit identified at time of analysis; EPSS is low (0.30%, 22nd percentile) and CISA SSVC records no observed exploitation.
Technical ContextAI
Fleet Server is the central control-plane component of the Elastic Stack that brokers communication between Elastic Agents and Kibana/Elasticsearch, handling agent enrollment, policy delivery, and artifact/diagnostic file uploads. The root cause is CWE-770 (Allocation of Resources Without Limits or Throttling): the upload handling path accepts and buffers request data without enforcing an upper bound on memory, so an oversized or malformed upload can be inflated into an unbounded in-memory allocation (CAPEC-130, Excessive Allocation). Affected CPE is cpe:2.3:a:elastic:fleet_server across the listed version ranges. Because the impact is memory exhaustion of the Go-based server process, a successful request can crash or hang the service rather than corrupting data.
RemediationAI
Vendor-released patch: upgrade Fleet Server to 8.19.11, 9.2.5, or 9.3.0 as documented in Elastic advisory ESA-2026-44 (https://discuss.elastic.co/t/fleet-server-8-19-11-9-2-5-9-3-0-security-update-esa-2026-44); 8.x users on 8.0.0-8.19.10 should move to 8.19.11, and 9.x users on 9.0.0-9.2.4 should move to 9.2.5 or 9.3.0. Until patching, reduce exposure by restricting network access to the Fleet Server upload/HTTP port so only trusted Elastic Agent subnets can reach it (via firewall or security group), and place Fleet Server behind a reverse proxy or load balancer that enforces request-body size limits and per-client rate limiting on upload endpoints - the trade-off is that overly strict body-size caps can break legitimate large agent diagnostics/artifact uploads, so tune limits to your fleet's real upload sizes. Monitor Fleet Server memory and configure process/container memory limits with automatic restart to blunt the impact of a successful allocation attack.
More in Fleet Server
View allAn issue was discovered in Fleet Server >= v8.10.0 and < v8.10.3 where Agent enrolment tokens are being inserted into th
Server-side code injection in Elastic Fleet Server (8.5.0-8.19.19, 9.0.0-9.4.4, and 9.5.0) lets an actor who can define
Policy information disclosure in Elastic Fleet Server allows any enrolled agent with valid credentials to retrieve polic
Incorrect authorization in Elastic Fleet Server allows any authenticated enrolled agent to interfere with multi-part upl
Same technique Denial Of Service
View allShare
External POC / Exploit Code
Leaving vuln.today