A mesh trust rotation succeeds in staging and later one production namespace alone cannot reach external TLS destinations through the egress gateway.
A service mesh egress gateway looks healthy and outbound TLS still fails
The gateway is updated, yet a subset of workloads still reject the path because their sidecars ship older trust material.
Scenario
What to check first
- Identify the primary failure signal in the The Gateway Learned the New Trust and One Namespace Stayed in the Old Epoch scenario.
- Separate visible symptoms from the underlying technical dependency.
- Describe the safest recovery path and the follow-up prevention work.
Checking checklist
- Summarize the current impact and the last known change.
- Collect direct evidence from logs, runtime state, and configuration before changing anything.
- Separate immediate recovery from permanent prevention work.
Recovery and prevention
Compare sidecar image trust bundles before rotating the gateway again or bypassing TLS verification.
Questions worth viewing together
Community-field Kubernetes problem inspired by service mesh and Kubernetes discussions where one namespace kept an older trust bundle after egress gatew... Mesh trust rotations can fail asymmetrically when some namespaces pin older sidecar images.
Teams often blame gateway cert deployment when the caller sidecar still trusts the old root only.
Trust rotations should include sidecar version and root bundle audits across every namespace, not just gateway pods.
Similar cases seen in the field