scaleway/pigeon-cluster v0.3.2
`compute-instance` v5.6.0 (#246) fixed the `Invalid count argument` error that broke `pigeon-cluster`'s private-network wiring on every real first apply, by replacing the implicit `private_network_id != null` gate with an explicit `enable_private_network` boolean.
This updates `pigeon-cluster`'s internal `module "job"` call to set `enable_private_network = true` (unconditionally — every job always attaches to the cluster's one shared Private Network) and bumps its pinned `compute-instance` source from `v5.4.0`/`v5.5.0` to the now-real `v5.6.0`. Confirmed via `tofu validate` against the real, published `v5.6.0` tag.
No public interface change to `pigeon-cluster` itself — purely an internal composition fix.