Add the ability to filter query args from the cache key to improve the cache HIT rate

There are certain query string arguments we want to ignore it terms of caching. For example, this URL https://theincline.com/2018/07/18/this-state-senator-wants-pennsylvania-to-ban-the-outdated-gay-and-trans-panic-defense/?mc_cid=5035471bd9&mc_eid=b31feb6b33 we can ignore the mc_cid and mc_eid query arguments since those args don't affect how the page is rendered.

To do this we need to preprocess the cache key. The SRCache nginx module even has a section in the README about preprocessing the cache key: https://github.com/openresty/srcache-nginx-module#cache-key-preprocessing

This is the nginx part of https://github.com/spiritedmedia/spiritedmedia/issues/2804

What query strings can we ignore?

Looking at Google Analytics can show us which query args are most frequently used:

refresh
mc_cid
mc_eid
fbclid
utm_source
utm_medium
utm_campaign

We definitely can't ignore the search query, s, as that modifies what is rendered on the page.

Update `/etc/nginx/sites-available/<domain-name>`

We need to extract the path from the $request_uri variable and set a new variable. nginx doesn't allow modifying the $request_uri variable directly (which makes total sense)

The <domain-name> value is different depending on the enviornment: spiritedmedia.dev, staging.spiritedmedia.com, spiritedmedia.com This code needs to be run in the http context.

# Extract the path from $request_uri and store it in a new variable $request_uri_path
# See https://stackoverflow.com/a/43749234/1119655
map $request_uri $request_uri_path {
  "~^(?P<path>[^?]*)(\?.*)?$"  $path;
}

Update `/etc/nginx/common/redis-php7-modified.conf`

This is the new configuration so we can ignore certain query strings.

# Modified Redis NGINX CONFIGURATION
set $skip_cache 0;
# POST requests and URL with a query string should always go to php
if ($request_method = POST) {
  set $skip_cache 1;
}

# We want to cache requests with query strings
#if ($query_string != "") {
#  set $skip_cache 1;
#}

# Don't cache URL containing the following segments
if ($request_uri ~* "(/wp-admin/|wp-.*.php|index.php|sitemap(_index)?.xml|[a-z0-9_-]+-sitemap([0-9]+)?.xml)") {
  set $skip_cache 1;
}

# Don't use the cache for logged in users or recent commenter or customer with items in cart
if ($http_cookie ~* "comment_author|wordpress_[a-f0-9]+|wp-postpass|wordpress_no_cache|wordpress_logged_in|woocommerce_items_in_cart") {
  set $skip_cache 1;
}

# Strip certain query args by setting them to null so they can be ignored in the cache key 
# Improves our cache HIT rate
rewrite_by_lua '
  local args = ngx.req.get_uri_args()
  args.refresh = nil
  args.mc_cid = nil
  args.mc_eid = nil
  args.fbclid = nil
  args.utm_source = nil
  args.utm_medium = nil
  args.utm_campaign = nil
  ngx.req.set_uri_args(args)
';

# Use cached or actual file if they exists, Otherwise pass request to WordPress
location / {
  try_files $uri $uri/ /index.php?$args;
}

location /redis-fetch {
    internal;
    set  $redis_key $args;
    redis_pass  redis;
}

location /redis-store {
    internal;
    set_unescape_uri $key $arg_key ;
    redis2_query  set $key $echo_request_body;
    redis2_query expire $key 14400;
    redis2_pass  redis;
}

location ~ \.php$ {
  # Force $request_uri_path to lowercase
  # See https://stackoverflow.com/a/48060609/1119655
  set_by_lua $request_uri_path "return ngx.arg[1]:lower()" $request_uri_path;  

  # $request_uri_path is defined in the http context in nginx.conf
  set $key "nginx-cache:$scheme$request_method$host$request_uri_path?$args";
  try_files $uri =404;

  srcache_fetch_skip $skip_cache;
  srcache_store_skip $skip_cache;

  srcache_response_cache_control off;

  set_escape_uri $escaped_key $key;

  srcache_fetch GET /redis-fetch $key;
  srcache_store PUT /redis-store key=$escaped_key;

  more_set_headers 'X-SRCache-Cache-Key $key';
  more_set_headers 'X-SRCache-Fetch-Status $srcache_fetch_status';
  more_set_headers 'X-SRCache-Store-Status $srcache_store_status';

  include fastcgi_params;
  fastcgi_pass php7;
}

spiritedmedia / systems