Index lifecycle capacity checks instead of scanning retained jobs

src/lib/server/hosts/schema.ts

@@ -44,6 +44,8 @@ export function ensureHostSchema(handle: Database.Database): void {
         -- causing a history scan for every queued command on Node's main thread.
         CREATE INDEX IF NOT EXISTS ec_host_operations_kind ON ec_host_operations(host_id, kind, created_at);
         CREATE INDEX IF NOT EXISTS ec_host_operations_kind_state ON ec_host_operations(kind, state);
+        CREATE INDEX IF NOT EXISTS ec_host_operations_active ON ec_host_operations(state, id)
+            WHERE state IN ('pending','accepted','running','unknown');
         CREATE TABLE IF NOT EXISTS ec_host_operation_payloads (
             operation_id TEXT PRIMARY KEY REFERENCES ec_host_operations(id) ON DELETE CASCADE,
             payload TEXT NOT NULL