For the complete documentation index, see llms.txt. Markdown versions of all docs pages are available by appending .md to any docs URL.

Solo build of Istio version 1.30.3-patch0 patch release.

This release note describes what’s different between Solo builds of Istio versions 1.30.3 and 1.30.3-patch0.

General Changes

• Built against upstream Istio commit da9c990125a8a69e59c44474a28470c9762c9e26. Compare.

Solo Flavor Changes

• Improved istioctl multicluster link to support a --reconcile option. When enabled, any “istio-remote” gateway in a “from” context that is not in the “to” context is deleted.

• Added the ability to specify the polling interval of ECS workloads discovery using the environment variable, AWS_POLLING_INTERVAL.

• Added a new environment variable, AWS_POLLING_INTERVAL, which will control the discovery of EC2 workloads. The default value is 1 minute.

• Added opt-in support for EnterpriseListenerSet (enterprise.solo.io/v1alpha1), a Solo equivalent of the Gateway API ListenerSet for platforms that pin Gateway API below 1.5. Enable with istiod environment variable SOLO_ENABLE_ENTERPRISE_LISTENER_SET=true (default false — no behavior change). A Gateway API ListenerSet takes precedence over an EnterpriseListenerSet wherever the two collide. Cross-namespace references from an EnterpriseListenerSet are authorized by a ReferenceGrant naming kind ListenerSet in group gateway.networking.k8s.io, so grants keep working unchanged when migrating to ListenerSet; a grant naming EnterpriseListenerSet is not recognized.

• Fixed generated east-west gateway regeneration when PEERING_AUTOMATIC_LOCAL_GATEWAY is enabled and a generated istio-remote gateway is manually modified or deleted.

• Fixed ztunnel to report the correct egress policies.

• Fixed ztunnel chart examples for egressPolicies.

• Fixed istioctl ecs add-service to handle empty docker labels when specifying ecs-ec2.

• Fixed istioctl ecs add-service to associate the ServiceAccount to task execution role when specifying ecs-ec2.

• Fixed an issue where, after a pod restart, its sidecar was missing the inbound routes for the ServiceEntry hostnames that select it. In multicluster peering, calls to that pod’s global <service>.<namespace>.mesh.internal hostname from another cluster failed with 503 UC, and the only recovery was to restart istiod.

• Fixed applying the draining weight from the east-west gateway when no cluster annotation existed.

FIPS Flavor Changes

No changes in this section.