Data & storage / 5 / 11
Sharding
A chat service's messages table reaches 6 TB and one server can no longer hold it. We split it across shards, choose a shard key, meet a hot shard and a hot key, and add a fourth shard without moving everything.
- 13 min
- Intermediate
- 10 steps
- Query p95
- 1 400 ms
- Hottest shard load
- 98 %
- Shards per query
- 1
- Largest shard
- 6 TB
- request
- reply
- redelivery
- duplicate
- lost
Numbers are a model scenario, not measurements
The real request
Chat API → Messages DBRequest
INSERT INTO messages (chat_id, sender_id, body)1
VALUES ('c_8812', 'u_42', 'see you at 7');
Response
INSERT 0 1
Time: 1380.5 ms2
chat_id— The chat this message belongs to. Almost every query filters on it.Time— 1.4 seconds to insert one message. The server is out of room, not the query.
Course: 42 lessons
01 Foundations
02 Scaling
03 Data & storage
04 Communication
05 Reliability
06 Architecture
07 Security
08 Shipping
09 Whole systems
10 Business processes