Event journal in Azure (no Cosmos inside)
summary
Implements event journals using Azure Table plus Blob storage with an HTTP enqueue -> queue-trigger -> writer pipeline (Azure Functions and Cloud Queue), avoiding CosmosDB and cutting estimated operational cost by about 10x. Uses entityId+sequenceNumber primary key, compresses payloads to fit the 64KB table field and spills oversized events to blob, and applies an inter-partition secondary index pattern to support time-range replay. Runs in production, supports audit-style replays after outages, but sacrifices built-in geo-replication and requires custom backup/retention.
AI-written summary · about
chapters (8)
- 0:00Introduction to event journals
- 4:00Why event journals in Azure
- 16:41Event sourcing basics
- 21:59Storage options in Azure
- 27:39Cosmos DB considerations
- 31:12Cost comparison table storage vs Cosmos
- 35:23Implementation examples
- 42:32Conclusion and Q&A
related talks
51mImplementing an Event Sourcing strategy on Azure
39mBuilding Event-Driven Systems at Scale with Cosmos DB
31mEvent-Driven Architecture - A successful strategy for a journey to the cloud
35mBuilding Event Driven Integration Architectures
20mHow Azure Cosmos DB enables Event Sourcing architecture
27mAzure Cosmos DB Changefeed tales from the trenches
03h 12mEMEA Live Stream
20mHow Azure Cosmos DB enables Event Sourcing architecture and powers e-commerce