Open leoluk opened 4 years ago
We don't need 1-second high resolution data for long-term archival.
Use an AggreatingMergeTree, aggregate it to 5m or so and possibly keep min/max/median.
This would require flows_raw to be migrated to the new table type OR we create a second table and use a view. If we do the queries correct there shouldnt be any overhead
We don't need 1-second high resolution data for long-term archival.
Use an AggreatingMergeTree, aggregate it to 5m or so and possibly keep min/max/median.