
Cloud Vulnerability DB
A community-led vulnerabilities database
CVE-2026-128085 is a denial-of-service vulnerability in OpenTelemetry-Go (opentelemetry-go) caused by an incomplete fix for a prior advisory (GHSA-p9f8-wvj8-2fg8). The AttributeValueLengthLimit and WithAttributeValueLengthLimit truncation logic in sdk/internal/attrnorm, sdk/log/internal/attrnorm, and trace/auto.go fails to count valid U+FFFD (REPLACEMENT CHARACTER) runes, allowing attribute values composed of those characters to bypass configured memory bounds entirely. Affected versions span SDK v1.33.0–v1.46.x, SDK log v0.9.0–v1.46.x, and trace v1.35.0–v1.46.x; all are fixed in v1.47.0. It carries a CVSS v3.1 base score of 5.3 (Medium) (GitHub Advisory).
The root cause is classified as CWE-135 (Incorrect Calculation of Multi-Byte String Length) and CWE-770 (Allocation of Resources Without Limits or Throttling). The shared truncate() helper (generated from internal/shared/attrnorm/truncate.go.tmpl) uses an "optimistic" fast path that iterates over the string with a range loop; when it encounters utf8.RuneError, it checks the decoded size to distinguish invalid bytes (size 1) from a legitimately encoded U+FFFD (size 3). Only the size-1 case diverts to the slow path — the size-3 U+FFFD case neither increments the character counter nor triggers truncation, so the rune is silently skipped. As a result, a 1,000-rune U+FFFD string (3,000 bytes) passes through completely untruncated even with a configured limit of 5 characters (15 bytes). The vulnerability is exploitable remotely without authentication, as span and log-record attributes are routinely derived from untrusted input such as HTTP headers or request bodies (GitHub Advisory, Fix Commit).
Successful exploitation allows an unauthenticated network attacker to inflate per-span and per-log-record memory retention linearly by injecting attribute values composed of U+FFFD runes, defeating the memory protection that operators explicitly configured via AttributeValueLengthLimit. Over time, this can exhaust memory and storage in observability pipelines (collectors, backends), degrading or denying service to the telemetry infrastructure. Confidentiality and integrity are not affected; the impact is limited to availability of the observability system (GitHub Advisory).
No public proof-of-concept exploit code has been published, and there is no evidence of in-the-wild exploitation at this time (GitHub Advisory). The EPSS score is 0.0, reflecting very low current exploitation probability. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities catalog. Exploitation requires only network access and no authentication, but the impact is limited to denial of service against observability components rather than the instrumented application itself.
AttributeValueLengthLimit configured and exposes an endpoint that reflects attacker-controlled data into telemetry attributes (e.g., HTTP headers, query parameters, request body fields).\uFFFD (3,000 bytes). This value will bypass the truncation limit regardless of the configured character limit.span.SetAttributes() or record.AddAttributes()).0xEF 0xBF 0xBD).\uFFFD sequences.\xEF\xBF\xBD byte sequences (U+FFFD in UTF-8) in headers, query strings, or body fields that are captured as telemetry attributes.Upgrade all affected OpenTelemetry-Go modules to version 1.47.0, which correctly counts valid U+FFFD runes toward the attribute value length limit: go.opentelemetry.io/otel/sdk v1.47.0, go.opentelemetry.io/otel/sdk/log v1.47.0, and go.opentelemetry.io/otel/trace v1.47.0 (SDK Release, SDK Log Release, Trace Release). As a short-term workaround prior to patching, implement input validation and rate limiting at the ingestion point to reject or truncate excessively large attribute values before they reach the SDK. Additionally, consider disabling AttributeValueLengthLimit enforcement reliance as a sole memory control and supplement with process-level memory limits on collector and backend components.
Fix availability across major Linux distributions and their releases.
Source: This report was generated using AI
Free Vulnerability Assessment
Evaluate your cloud security practices across 9 security domains to benchmark your risk level and identify gaps in your defenses.
Get a personalized demo
"Best User Experience I have ever seen, provides full visibility to cloud workloads."
"Wiz provides a single pane of glass to see what is going on in our cloud environments."
"We know that if Wiz identifies something as critical, it actually is."