C4 model of a messaging API, all four levels

Views in this map

1. 1
   L1 ContextWho uses Gridline, and the carriers, providers and services it relies on.
2. 2
   L2 ContainersTen containers: the public API, the send path, webhooks and analytics, and their stores.
3. 3
   L3 Rate limiterInside the rate limiter: rules, overrides and a token bucket per rule, kept in Redis.
4. 4
   L4 Token bucketThe token bucket’s code: TokenBucket, the store it depends on, and the Lua Redis runs.

L1 Context ▸ L2 Containers ▸ L3 Rate limiter ▸ L4 Token bucket

Flows to play

• ▷
  One SMS requestFrom the customer’s POST to the carrier, through the rate limiter. In L2 Containers: the customer’s app, the REST API, the rate limiter, the router, a delivery worker and the carrier, a step each.
• ▷
  A rate-limit checkOne Check() call through the rate limiter’s components. In L3 Rate limiter: the same check inside the rate limiter, from the Limit API through the rules to Redis and back.
• ▷
  tryConsume()One call into the token bucket’s code, down to Redis and back. In L4 Token bucket: the call through TokenBucket, its clock, rule and store, to the Lua script that Redis runs.

The model behind it

57 records of 7 types. Every box on the map is one of them, so the filters and the record panel read the same data.

• Components 26
• Containers 10
• Code elements 7
• External systems 6
• Teams 5
• People 2
• Software system 1

Filters

Level and Language: pick a value and everything else fades.