Summary
Istio's JWT authentication logic could fall back to hardcoded default behavior when the JWKS resolver is unavailable or JWKS fetching fails, potentially ignoring intended RequestAuthentication policies.
Under these failure conditions, services that are assumed to be protected by JWT validation may instead accept requests without proper verification, exposing data or functionality that operators believed was access-controlled.
Affected context
This vulnerability affects Istio deployments that rely on JWT-based authentication, particularly those using RequestAuthentication resources to validate tokens against remote JWKS endpoints.
It manifests when the JWKS resolver becomes unavailable (e.g., network partitions, DNS issues, endpoint downtime) or when JWKS fetching consistently fails.
In these cases, instead of enforcing the configured authentication policy, the control plane and/or proxies can fall back to hardcoded defaults, weakening or bypassing expected token verification and leaving services less protected than operators intend.
Exploit preconditions
- Attack surface: HTTP/gRPC endpoints exposed through the Istio service mesh and protected via JWT-based policies.
- Required access: Ability to send requests to mesh workloads that are assumed to be JWT-protected; in some scenarios, influence over network conditions or JWKS availability increases risk.
- Trigger point: JWKS resolver unavailability or repeated failure to fetch JWKS data (e.g., unreachable issuer, timeouts, TLS errors, or misconfiguration).
- Why exploitation is possible: When JWKS retrieval fails, the system may revert to hardcoded default behavior rather than failing closed or consistently enforcing the configured RequestAuthentication rules. This error handling pattern can cause authentication checks to be skipped or applied less strictly, exposing resources that should require valid tokens.
Why this happens
The underlying issue stems from error handling and default configuration behavior around JWT key resolution.
Instead of treating JWKS fetch failures as hard failures that preserve security guarantees, the implementation relies on fallback defaults that do not accurately reflect the operator's RequestAuthentication configuration.
This aligns with CWE-200: Information Exposure, where sensitive resources become accessible because a component reveals or permits access under error conditions, and with a broader pattern of "fail-open" authentication logic where validation failure does not reliably block access.
Potential impact
- Services that operators believe are protected by JWT authentication may accept unauthenticated or improperly validated requests.
- Attackers could access APIs or data that are intended to be restricted to authenticated users or specific identities.
- Lateral movement within the mesh is facilitated if internal services rely on JWT-based trust assumptions that are weakened by fallback behavior.
- Compliance or policy violations if regulated data is served without the expected authentication controls.
General mitigation guidance
Resilience against this class of issue generally focuses on fail-closed authentication and robust treatment of remote key dependencies:
- Design authentication logic to treat JWKS retrieval failures as security events, defaulting to denial rather than permissive access when keys cannot be validated.
- Reduce reliance on remote-only JWKS sources where feasible by caching keys securely and validating cache freshness, while still failing closed if keys cannot be confirmed.
- Harden mesh configuration so that services that must be protected by JWTs also enforce complementary checks (e.g., authorization policies) that do not silently weaken on upstream failures.
- Clearly separate "optional" authentication from "mandatory" authentication in configuration so that failure modes are explicit and conservative.
- Apply least privilege and data minimization at the service layer so that even if authentication enforcement degrades, exposed endpoints reveal as little sensitive information as possible.
How teams detect issues like this
Detection typically involves validating both configuration intent and runtime behavior around JWT enforcement:
- Use configuration review or policy-as-code to verify that critical services have strict RequestAuthentication and AuthorizationPolicy rules, and that these rules are not defined in a way that tolerates silent fallback.
- Employ static or configuration analysis tools that flag authentication policies dependent on external JWKS endpoints without explicit fail-closed semantics.
- Monitor mesh and control-plane metrics/logs for JWKS fetch failures, resolver unavailability, or frequent key-refresh errors, and correlate these with access logs to ensure requests are not being allowed when validation cannot be performed.
- Add security tests or chaos experiments that simulate JWKS unavailability and verify that protected endpoints remain inaccessible without valid tokens.

