Cloud Run VPC access has max instance constraint
Previously, AI Gateway configured spec.scalability.max_instances arbitrarily to 200. To learn more, refer to https://cloud.google.com/run/docs/configuring/max-instances.
In gitlab-org/modelops/applied-ml/code-suggestions/ai-assist!1836 (merged), VPC access was enabled in staging. Cloud Run has a constraint that resulted in the following error from https://gitlab.com/gitlab-com/gl-infra/platform/runway/deployments/ai-gateway/-/jobs/9012193751:
[
{
"@type": "type.googleapis.com/google.rpc.BadRequest",
"fieldViolations": [
{
"description": "Invalid value 200. autoscaling.knative.dev/maxScale annotation must be less than or equal to 100 when run.googleapis.com/network-interfaces annotation is set.",
"field": "service.template.scaling.max_instance_count"
}
]
}
]
Update Runway service manifest validation to handle edge case due to Cloud Run constraint.