Description:
When setting metrics.clusterStatName for a EnvoyProxy resource, an incorrectly formatted value (missing a % symbol) caused Envoy Gateway to delete and recreate the Envoy Proxy Deployment with a different name, resulting in traffic disruption.

For example:

Incorrect (missing % after ROUTE_RULE_NAME):

metrics:
    clusterStatName: '%ROUTE_KIND%/%ROUTE_NAMESPACE%/%ROUTE_NAME%/rule/%ROUTE_RULE_NAME/%ROUTE_RULE_NUMBER%'

Correct:

metrics:
    clusterStatName:  '%ROUTE_KIND%/%ROUTE_NAMESPACE%/%ROUTE_NAME%/rule/%ROUTE_RULE_NAME%/%ROUTE_RULE_NUMBER%'

After applying the incorrect format, the existing Envoy Proxy Deployment (envoy-proxy-private) was deleted and replaced with a new Deployment named envoy-kube-system-private. All xRoutes failed until I corrected the typo, which triggered yet another redeployment to restore the original Deployment name.

Expected behavior:
A format error in metrics.clusterStatName should raise a warning in the logs.

Repro steps:

Deploy Envoy Gateway with an EnvoyProxy resource using the incorrect format for metrics.clusterStatName (missing a % symbol as shown above).

Apply the change and observe the name of the Envoy Proxy Deployment.

Fix the format string and apply again.

Expected:
The Deployment name remains the same, and Envoy is updated in-place.

Actual:
Deployment is deleted and recreated with a different name, causing traffic disruption.

Environment:

Envoy Gateway version: 1.5.0

Kubernetes version: 1.32

Logs:
N/A — issue observed from Kubernetes resource deletion/recreation