Time Series with SQL Server 2022
summary
Compares SQL Server 2022 to InfluxDB and shows concrete numbers and techniques: loads 15 days of performance counters (two instances, 82 million rows), finds heap/clustered index size ~26–28 GB versus clustered columnstore 400 MB (archive 231 MB) and notes InfluxDB v1 at 91 MB; partitioning for retention inflates size to ~2.6 GB; retention-policy syntax from SQL Edge is missing in 2022 so deletes require partition switching; reproduces Group-By-Time and fill using date_bucket + generate_series, and concludes SQL Server can work but needs manual effort.
AI-written summary · about
chapters (12)
- 0:00Introduction to time series
- 2:33Defining time series data
- 7:07Relational vs time series
- 10:45The importance of time series
- 12:30SQL Server for time series
- 15:00Introduction to InfluxDB
- 20:17InfluxDB design principles
- 26:19Storage optimization in SQL
- 33:54Retention and partitioning
- 40:23Partitioning demonstration
- 42:11Querying time series data
- 48:33Conclusion and Q&A
related talks
51mNext Level Scalability on SQL Server 2022
49mPostgreSQL, The Time-Series Database You Actually Want
23mTime series database: Should I use one?
39mA New Data Model for Time Series Data Processing
29mCatching Waves with Time-Series Data
49mHidden Gems in SQL Server 2022 Database Engine
21mSQL Server 2022 – Time to Rethink your Backup and Recovery Strategy
51mPostgreSQL, The Time-Series Database You Want