http.Fetch not getting latest results

Hello, I recently updated my script to load it’s config from a webpage so I don’t have to update it on all my different servers when I make changes. It seems to work fine except for when I refresh it. When I refresh it doesn’t get the latest result. It will say that it has worked even if I delete the config file from the webserver. It’s almost as if http.Fetch has a cache of some sort.

plt_config.lua (in autorun)

plt_conf_file = ""
function plt_conf_overrides()
	plt_cooldowntime = 20
	print("[Porchlight] Loaded override config successfully.")

plt_main.lua (in autorun)

function plt_conf_initf(reloaded)
	print("[Porchlight] Starting loading config file \"" .. plt_conf_file .. "\".")
	http.Fetch( plt_conf_file,
		function( body, len, headers, code )
			print("[Porchlight] HTTP CODE : " .. code)
			if code == 200 then
				RunString( body, "plt_conf_loader", true )
				if reloaded then
					print("[Porchlight] Config file reloaded successfully.")
				print("[Porchlight] ERROR : Config loaded unsuccessfully.")
		function( error )
			print("[Porchlight] ERROR : Config loaded unsuccessfully.")
hook.Add( "Initialize", "plt_conf_inith", plt_conf_initf(false) )

Running plt_conf_initf() again will return

[Porchlight] Starting loading config file "".
[Porchlight] HTTP CODE : 200
[Porchlight] Loaded base config successfully.
[Porchlight] Loaded override config successfully.

even if config.lua doesn’t exist on the webserver. Is it possible that it does this because the config file is so large? Is there a better way I should be doing this? Or if there is a cache, how can I clear it every time the function runs?

Okay, so I’ve made myself a workaround.

plt_conf_file_mod = plt_conf_file .. "?" .. tostring(math.random( 1, 99999 )) .. "=" .. tostring(math.random( 1, 99999 ))

I redefine the url so it creates random meaningless php vars like config.lua?91571=21238, and it works. However, this solution feels messy and imperfect.

This makes me more confident in my cache theory. There has to be a way just to clear the stored page, right?

Ya sure there isn’t a CDN like cloudflare doing all the caching? A completely lazy way to guarantee no caching do it is to use POST instead of GET.

just use os.time() bro

Edit your .htaccess to prevent .lua caching

<filesMatch "\.(html|htm|js|css)$">
  FileETag None
  <ifModule mod_headers.c>
     Header unset ETag
     Header set Cache-Control "max-age=0, no-cache, no-store, must-revalidate"
     Header set Pragma "no-cache"
     Header set Expires "Wed, 11 Jan 1984 05:00:00 GMT"

Thank you. That fixed it. Needed to enable mod_headers however.