Given dfs['events'] (event_id, event_ts, amount), protect downstream state from late retries and unbounded memory growth:
.withWatermark('event_ts', '15 minutes').event_ts >= '2026-05-01 11:45:00')..dropDuplicates(['event_id']).event_id, event_ts, amount ordered by event_id.These DataFrames are passed to solve(spark, dfs). Access one with dfs['name'].