computeplane: the module release can time out and uninstall the cluster it is still waiting for #4714

The computeplane HelmRelease waits on the computeplane-cluster release it renders, but has only a 10m install timeout and unlimited remediation retries. A tenant cluster bring-up routinely takes longer than 10 minutes (worker join alone has a 29m budget in the kubernetes e2e suites). When the outer timeout fires first, Flux remediates by uninstalling and reinstalling the module, which tears down the cluster release mid-bringup, and the cycle can repeat. Found by reading the config while writing the computeplane e2e suite (#4713), not yet seen on a cluster; the suite's status.history remediation guard would catch it. Fix shape: give the module release a timeout above the cluster release's own (30m) plus margin, or stop it from waiting on the inner release, and keep remediation from uninstalling a release that is still installing.