Only one proxy is being used, or is that intentional?

[zunjae]

Maybe I am misunderstanding how this project works...

I loaded in my webshare link which has 250 proxies. For the past hour, performing a few hundred to a thousand API calls, all my call have used the same proxy address. I would like it to use random ones. Your readme does imply that is possible

"A proxy server that distributes your requests over a pool of proxies"

and

"It makes individual requests over random proxies from the pool"

---

My config:

  proxymanager:
    container_name: proxymanager
    image: ghcr.io/jkelin/forward-proxy-manager:latest
    environment:
      - PROXY_LIST_URL=webshare_link_here

In my API I use http://proxymanager:8080. I've tested with https://api.ipify.org?format=json to test if I'm indeed using the proxy server and I am!

What am I doing wrong? Why is it not picking other proxies?

[jkelin]

It is possible, it's the whole point of the application in fact.

During application startup it should test all of the proxies for liveness and you will see all of them inside terminal like this, do you see all of your proxies?

What kind of client do you use?

Thinking more about it, is it possible that you are not actually hitting real or virtual api limits because you are requesting the api slowly enough? For example with default configuration with 1000 request sample, the requests are evenly distributed over 100 proxies, but if there is a 1 second delay between requests, I'm only seeing request distribution over two proxies because limits are replenished quicker than they are exhausted.

[jkelin]

I've added a patch for the issue I've outlined above. Proxies are now ordered randomly for each request even for low frequency usage.

Please try updating the image with docker pull ghcr.io/jkelin/forward-proxy-manager:latest and restarting your docker compose docker compose restart ..

[zunjae]

I'm now seeing a different IP per request, thank you!

but if there is a 1 second delay between requests, I'm only seeing request distribution over two proxies because limits are replenished quicker than they are exhausted.

Is it possible to still use the older approach where it keeps using 1 IP for a website in case I ever need it, preferably configurable through env variables?