Contextualizing timeseries data with graphs
summary
Kusto (Azure Data Explorer) adds in-query graph semantics that build an in-memory property graph per query using make graph and traverse it with graph match, combining time-series operations and anomaly detection in one statement. Demo scales to millions of nodes (median graph creation ~0 ms; 95th percentile ≈100 ms) and handles production use cases—co‑occurrence recommendations, lineage visuals, and blast‑radius impact—served from a service ingesting ~250 petabytes and billions of daily queries.
AI-written summary · about
chapters (14)
- 0:00Welcome
- 1:10Kusto introduction
- 5:21Contextualize time series with graphs
- 6:50Graphs and Kusto
- 9:16Graph features in Kusto
- 13:28Graph queries demo
- 18:52Time series graph integration
- 24:41Industrial IoT graphs
- 27:07Security graph semantics
- 28:51Graph lineage in Kusto
- 32:51Graph performance
- 36:00Next steps and GA plans
- 37:25Resources
- 38:54Q&A
related talks
51mAzure Data Explorer - a Torch in the Dungeon of Observational Analytics
42mThe Most Powerful Azure Service You’ve Never Heard Of
35mBuilding Real-Time Graph Analytics Platforms
57mGraphing Your Way Through the Cosmos: Common Data Problems Solved with Graphs
29mAll about graphs in Database
22mVisualizing Azure Cosmos DB with Linkurious
30mLarge Scale Graph Solutions: Use-cases and Lessons Learnt
49mModern Graph Databases