The map
- Not started
- In progress
- Done
- Builds on
- 01Documents and schemasShape data by how you read it0/3
- 02Finding things fastPlans, indexes, pipelines0/3
- 03Writes you can trustReplicas, acknowledgements, transactions0/3
- 04Scale and changeShards, streams, storage0/3
- 05Next: system designSame ideas, at service scale
- Read replicasLibrary ↗Reading from a secondary is reading from an asynchronous replica. The library topic covers stale reads and read-your-writes routing in general.
- ShardingLibrary ↗A shard key is a partitioning decision you live with for years. The library topic compares hash and range partitioning across systems.
- Change data captureLibrary ↗Change streams are MongoDB's built-in change data capture over the oplog. The library topic shows the pattern with other databases and Kafka.
- Transactional outboxLibrary ↗Write the order and its event in one transaction (or one document), then publish from a change stream, so you never commit one without the other.
- Leader electionLibrary ↗A replica set election is leader election with terms and majorities. The library topic explains why a majority is what prevents two leaders.
Written for MongoDB 7.0 and 8.x; the links point to the current manual. The examples were run on MongoDB 8.3.13 with mongosh 2.13.0, on a three-member replica set in Docker with 200,000 synthetic orders, so counts and timings will differ on your machine. Sharding examples were not run. MongoDB 9.0 was released in September 2026; where it changes something covered here, the topic says so.
On licensing: MongoDB Community Server versions released on or after 16 October 2018 are under the Server Side Public License (SSPL), a licence written by MongoDB Inc. and based on the GPLv3. It is not approved by the Open Source Initiative, which is why MongoDB is usually described as source-available rather than open source. Its extra copyleft condition (section 13) applies when you offer MongoDB's own functionality to third parties as a service; MongoDB's FAQ says other applications that merely use MongoDB, including SaaS products, are not subject to it. Earlier versions remain under the AGPLv3, and the official drivers are Apache 2.0. If licensing matters for your use, read the licence and its FAQ rather than a summary.
Where this path leads
- Scaling out →Spread load across many machines and keep them from drowning.
- Inside the database →How databases stay durable, scale reads, and share their changes.
- Kafka: the log at the centre of the system →How an append-only, partitioned, replicated log works, what producers and consumer groups really guarantee, what exactly-once does and does not cover, and the operational traps (lag, rebalances, too many partitions) that come with running one.