This is a public summary of the project's phased intent — not a committed
timeline. Phases already shipped stay listed so the arc is visible; later phases
are invitations to open an issue or PR
if one of them matters to you sooner.
Additional strategies driven by real migrations (non-string enum remap, quantity/duration helpers, optional CEL, …). The Strategy enum and discriminated ConversionRule union were built so each addition stays contained.
Deeper Crossplane packaging/metadata pointers and a storage-version migration playbook. Crossplane v1 vs v2 is not an open product decision: v2 is targeted and tested in CI; v1 is handled identically by the code path but untested in CI (Limitations) — this phase closes the CI-coverage gap rather than choosing a stance.
Benchmark suite (compile vs schema size, Convert latency vs array length), registry Set scaling, synthetic large ConversionReview batches — numbers that backfill Capacity planning.
Documented CI patterns for convctl test/diff across many kubecontexts, GitOps examples (Flux/Argo). Cross-cluster failover of conversion state remains an explicit non-goal.
Strategies stay additive. A new strategy is a *Params type, an Op, a compile resolver, webhook validation, a CLI fixture, and a docs page — see Adding a strategy.
Observability is chart-optional. ServiceMonitor / PrometheusRule / Grafana dashboards ship with the chart and stay off unless enabled.