Choosing an Azure Data Platform: PostgreSQL, MySQL, Redis, Synapse, Data Lake, Data Factory and Fabric

Microsoft Azure Architect Technologies AZ 300 55 min min read Updated: Nov 17, 2026 Intermediate
Choosing an Azure Data Platform: PostgreSQL, MySQL, Redis, Synapse, Data Lake, Data Factory and Fabric
Intermediate Topic 3 of 3

Open Source Databases

  • Azure Database for PostgreSQL – Flexible Server and MySQL – Flexible Server: managed, with zone-redundant HA, automated backups, read replicas, VNet integration or private endpoint, and stop/start to save cost
  • Use them to move open-source apps to the cloud with little change
bash az postgres flexible-server create -g rg-data -n pg-shop-482 -l centralindia --tier GeneralPurpose --sku-name Standard_D2ds_v5 \ --version 16 --high-availability ZoneRedundant --storage-size 128 --admin-user pgadmin --admin-password '<StrongPassword>'

Azure Cache for Redis

  • A managed in-memory cache for sessions, leaderboards, rate limits and caching database results
  • Pattern: cache-aside. Read from cache, on a miss read the database and fill the cache with a TTL
  • Premium and Enterprise tiers add clustering, persistence, zone redundancy and geo-replication
  • Check the current offering (Azure Managed Redis is the newer option) before deciding

Analytics Services

ServiceWhat it is
Azure Data Lake Storage Gen2Cheap, scalable lake on Blob with hierarchical namespace
Azure Data FactoryOrchestrate and move data (ETL/ELT) with pipelines and 90+ connectors
Azure Synapse AnalyticsData warehouse (dedicated SQL pool), serverless SQL, Spark, pipelines
Microsoft FabricSaaS analytics platform: OneLake, lakehouse, warehouse, pipelines and Power BI in one product
Azure DatabricksApache Spark based platform for data engineering and machine learning
Azure Stream Analytics / Event HubsReal-time streaming ingestion and queries
Azure Data ExplorerFast analytics on logs and time-series
Power BIDashboards and reports
Note Microsoft is investing heavily in Fabric. Check the current guidance for new analytics projects, because product direction changes.

The Big Decision Table

RequirementService
Relational app, minimal adminAzure SQL Database
SQL Server with Agent and linked serversSQL Managed Instance
Open-source relationalPostgreSQL / MySQL Flexible Server
Global JSON documents, millisecond latencyCosmos DB (NoSQL)
Cache and sessionsRedis
Images, files, backupsBlob Storage
Big data lakeData Lake Storage Gen2
Enterprise data warehouseSynapse dedicated SQL pool or Fabric Warehouse
Streaming eventsEvent Hubs + Stream Analytics
Simple key-value at low costTable Storage
Full-text searchAzure AI Search
Graph relationshipsCosmos DB Gremlin

Reference Architecture: Modern Data Platform

  1. Sources (apps, databases, files) are ingested by Data Factory or Event Hubs
  2. Raw data lands in Data Lake Storage (bronze), then is cleaned (silver) and modeled (gold) with Spark or SQL
  3. A warehouse or lakehouse serves reports
  4. Power BI shows dashboards
  5. Purview catalogs and governs data

Security for Data Services

  • Use private endpoints, Entra ID auth and managed identities
  • Use customer-managed keys where policy requires them
  • Turn on auditing and Defender plans
  • Mask or classify sensitive columns

Practice

  1. Choose a service for each row of the decision table without looking
  2. Draw the cache-aside flow for a product catalog
  3. Explain why Fabric and Synapse may both appear in an architecture

Conclusion

Pick the simplest managed service that meets the requirement: SQL for relational, Cosmos DB for global NoSQL, Redis for cache, Blob and Data Lake for files and big data, and Data Factory, Synapse or Fabric for analytics.

Quick Summary: open source databases; Redis; analytics; decision table; reference architecture.

Get Newsletter

Subscibe to our newsletter and we will notify you about the newest updates on Edugators