---
title: Cache
description: >-
  Super can be configured to use a cache to improve the performance of your
  proxy server.
---
> ## Documentation Index
> Fetch the complete documentation index at: https://docs.page/luizfonseca/super/llms.txt
> Use this file to discover all available pages before exploring further.

Super can be configured to use a cache to improve the performance of your proxy server.

## Cache configuration

The cache configuration is defined in the `routes` section of the Super configuration file.

Each route can have a `cache` section that specifies the cache configuration for that route.

The cache configuration includes the following options:

- `enabled`: Whether the cache is enabled for the route. Defaults to `false`.
- `cache_type`: Which cache backend to use. Defaults to `memcache`. Other options are `disk`.
- `expires_in_secs`: The number of seconds the cache should be valid for. Defaults to `360`.
- `stale_if_error_secs`: The number of seconds the cache should be valid for if an error occurs. Defaults to `60`.
- `stale_while_revalidate_secs`: The number of seconds the cache should be valid for if the response is revalidated. Defaults to `60`.
- `path`: The path to the cache directory. Defaults to `/tmp`.

Here's an example of a route with a cache configuration:

```hcl
# config.hcl file
# You might need to edit your /etc/hosts file here.
host "example.com" {
  location "/" {
    cache {
      enabled = true
      cache_type = "memcache"
      expires_in_secs = 360
      stale_if_error_secs = 60
      stale_while_revalidate_secs = 60
      path = "/tmp"
    }

    upstream "10.0.1.3:3000" {}
  }
}
```

## Memory limit (disk cache)

Responses served from the `disk` cache are also kept in memory, so that frequently requested responses don't need to be read from disk. The memory used for this is limited by `cache.memory_limit_mb` (shared by all routes, defaults to `256`). When the limit is reached, the least popular responses are only served from disk. Responses larger than 64MB are always served from disk.

```hcl
# config.hcl file
cache {
  memory_limit_mb = 256 # 0 disables keeping responses in memory
}
```

The limit applies to the cached responses themselves; the process will use somewhat more memory than this.

## Cache usage

When a request is made to a route with a cache configuration, Super will check if the response is already in the cache. If it is, the response will be served from the cache instead of making a new request to the upstream server.

If the response is not in the cache, Super will make a new request to the upstream server and cache the response. The cache will be updated with the new response if the response is valid for the configured expiration time.