Data stores
Every data technology the estate runs on, not just the shared SQL estate: MongoDB carries a large share of the cloud services, Redis is a second coupling plane, and DynamoDB and S3 underpin the payments and document platforms. Repos sharing a store are coupled even without a code edge. The three core SQL databases are binding points; stage 4 of the programme gives them named owners.
Repos using each technology
A repo can use several technologies. Click a bar to filter the stores below.
The shape of the data estate
SQL Server: 98 stores across 89 repos. The legacy centre of gravity, and where the shared-schema coupling lives.
MongoDB: 75 stores across 66 repos. The default operational store for the modern and cloud services, mostly one database per service, which is exactly the decoupled pattern the programme wants.
Redis: 11 stores across 46 repos. Caching, session state and streams; shared instances couple services as a second plane.
Amazon S3: 13 stores across 12 repos. Object storage for documents, assets and request logs.
DynamoDB: 8 stores across 6 repos. The payments platform's serverless store.
Azure Blob Storage: 1 store across 1 repo.
OpenSearch: 1 store across 1 repo.
Snowflake: 1 store across 1 repo.
Shared stores (2) — Amazon S3
| Store | Technology | Coupled repos | |
|---|---|---|---|
| s3:mail-message-data | Amazon S3 | 2 | tc-mail-composer, tc-mail-orchestrator |
| s3:Public_BlogVideos | Amazon S3 | 2 | asset-management, tc-public-assets-api |
Single-owner stores (11) — Amazon S3
Owned by one repo each; no cross-system coupling. This is the healthy pattern most of the MongoDB estate follows.