Skip to main content
Experimental feature. Response caching is in alpha and is subject to change, and should not be used in production environments.
The router caches the responses to federation entity fetches. An entity fetch is the request the router sends to a subgraph through the _entities root field to resolve fields of an entity that another subgraph owns. Root field fetches are not cached. Entries are keyed per entity and per selection set. Asking for different fields of the same entity does not share an entry. Response caching is disabled by default.

​
What gets cached

A subgraph response is cached only when its Cache-Control header asks to be. The rules apply in this order:
  1. no-store refuses caching.
  2. no-cache or private refuses caching, in any form. This includes the qualified no-cache="Set-Cookie" form.
  3. public is required. Cache-Control: max-age=60 on its own is not cached.
  4. s-maxage sets the lifetime and takes precedence over max-age.
  5. max-age sets the lifetime when s-maxage is absent.
  6. fallback_ttl sets the lifetime when the response says public and names no lifetime of its own.
A value of 0 for s-maxage or max-age refuses caching. It does not fall through to the next rule. A subgraph that sends no Cache-Control header at all is never cached. fallback_ttl does not apply to it. A response carrying GraphQL errors is never cached, whatever its Cache-Control header says.
Configuring the response cache does not make anything cacheable on its own. Subgraphs opt in by sending Cache-Control: public. If nothing appears to be cached, check the subgraph response headers first.

​
Enable it

​
Redis

Every router replica shares one cache and entries outlive the process. Define a storage provider, then reference it by provider_id.
An unknown provider_id fails startup.

​
Router memory

Each replica holds its own cache, nothing survives a restart, and no Redis is needed.
provider defaults to redis, so caching in memory has to be asked for by name. This prevents a missing provider_id from quietly turning one shared cache into one cache per replica.

​
Configuration reference

config.yaml
storage is required when enabled is true. Environment variables bypass the config schema validation. A fallback_ttl of zero or less fails startup either way.

​
Limitations

These apply to the current alpha. A batch is served from the cache only when every entity in it is present. One miss sends the whole batch to the subgraph, including the entities that were already cached. Entities are stored individually, so a later batch made up only of cached entities is a full hit. Cache hits produce no subgraph telemetry. A hit skips the subgraph fetch, so no subgraph span or metric is recorded for it. Subgraph request counts fall as the hit rate rises. Use them to confirm the cache is working, not to measure traffic. enable_multi_fetch disables response caching. When engine.enable_multi_fetch is on, the router merges entity fetches to the same subgraph within one parallel wave into a single request. Merged fetches are not cached, and nothing reports that they were skipped. Do not enable both and expect response caching.

​
Observability

The router logs once at startup when the cache is enabled. With the Redis provider:
With the memory provider:
Nothing is logged when the cache is disabled. Cache failures never fail a request. When a read or a write fails, the router serves from the subgraph and logs a warning:
That warning is sampled to one line per second. An unreachable cache produces one failure per entity fetch of every request in flight, and logging it unsampled would compound the outage.
  • Storage Providers defines the Redis instance the redis provider references.
  • Cache Warmer is a different cache. It pre-populates the operation plan cache and does not cache subgraph responses.