Sporadic connection issues when using Nginx to forward Telegram Bot API requests

Other Questions Telegram bot
1

I’m having trouble with my Nginx setup for a Telegram Bot. It works fine when I just use my local server, but adding Telegram’s server causes random timeouts. Here’s what I’ve got:

upstream bot_servers {
    server localhost:8443;
    server bot.telegram.org:443;
}

server {
    listen 8080;
    server_name mybot.example.com;

    location /bot {
        proxy_pass https://bot_servers;
        proxy_set_header Host $host;
        proxy_ssl_name bot.telegram.org;
        proxy_ssl_protocols TLSv1.2 TLSv1.3;
        proxy_connect_timeout 30s;
        proxy_read_timeout 60s;
    }
}

When I use just my local server, everything’s smooth. But with Telegram’s server in the mix, I get random timeouts. The logs show:

[error] 12345#12345: *1 upstream timed out (110: Connection timed out) while connecting to upstream, client: 1.2.3.4, server: mybot.example.com, request: "POST /bot123456:ABC-DEF1234ghIkl-zyx57W2v1u123ew11/getMe HTTP/1.1", upstream: "https://149.154.167.220:443/bot123456:ABC-DEF1234ghIkl-zyx57W2v1u123ew11/getMe", host: "mybot.example.com"

I’ve checked my network, tweaked timeouts, but can’t figure out why it’s acting up. Any ideas on fixing these random timeouts when using Telegram’s API?

2

I’ve encountered similar challenges with Nginx and Telegram Bot API. One potential solution is to implement a health check for your upstream servers. This can help Nginx route traffic more efficiently and avoid timeouts. You might want to add something like this to your configuration:

upstream bot_servers {
    server localhost:8443 max_fails=3 fail_timeout=30s;
    server bot.telegram.org:443 backup;
    check interval=5000 rise=2 fall=3 timeout=4000;
}

This setup designates the Telegram server as a backup and implements a health check mechanism. It could potentially mitigate the sporadic connection issues you’re experiencing. Additionally, consider logging more detailed information about the requests to help diagnose the root cause of the timeouts.

3

I can relate to the frustration with Nginx and its unpredictable behavior when interfacing with external APIs like Telegram. In my experience, integrating a caching mechanism has helped reduce these sporadic timeouts. The idea is to cache successful responses for a short duration and also allow for a few retries when errors occur, which gives your setup a better chance of recovering from temporary glitches. This approach helped stabilize my connections and improve performance, so it might be a good strategy to consider for your configuration as well.

4

hey, i’ve had similar issues. have u tried adjusting ur keepalive settings? sometimes that helps with sporadic timeouts. also, maybe try using a single upstream server and handle failover in ur app instead. nginx can be finicky with multiple upstreams for stuff like this