Cache
Microlink API has a built-in cache layer to speed up consecutive requests over the same resource.
The cache layer is enabled for any request under the following workflow:
- The first time a resource is requested, a cache copy will be created.
- Sucessive access over the resource will consume the cache copy.
The cache layer is a combination of a unified cache and edge node cache. The combination provides
eventual consistency
.Unified cache
When you query for a resource against Microlink API for the first time, the request will generate a shallow cache copy.
That known as 'MISS' and it's reflect as
x-cache-status on response headers.Any successive API access based on the same URL will consume the shallow copy created, reflecting a 'HIT' at
x-cache-status response headers.Edge node cache
Since Microlink relies on
CloudFlare CDN
, after the unified cache is warm, any successive API access based on the sam URL will be served from the nearest edge node over CloudFlare Network
.That means not only response will be served from cache, also it will be served from the nearest distance relative to the request origin. That's reflected under
cf-cache-status response header as 'HIT'.Edge nodes cache is per edge location, meaning every edge node as their own cache, causing a 'MISS' reflected at
cf-cache-status response header when the access comes from a different location.When this happen, the edge node cache will fallback automatically into the unified cache, creating a new edge cache copy.
Cache invalidation
The cached request will be considered as valid until it reached the expiration time, reflected at the cache-control response header.
There is two ways to setup the expiration: