# Real-time alerts on SSH logins, and what do you do when one looks wrong? — retained record

> Retained reddit record from reddit:selfhosted.

Canonical: https://highsignal.app/data/records/0456f0c3a54adbd9

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:selfhosted)  published 2026-09-18 16:13 ingested 2026-09-19 02:39

[original source ↗](https://www.reddit.com/r/selfhosted/comments/1wjuk4k/realtime_alerts_on_ssh_logins_and_what_do_you_do/)

## retained text



```
A few non-technical people on our team upload files over SFTP with FileZilla, so password auth has to stay on for now. I know, keys only, FileZilla supports keys, a VPN would be better. It's on the list. fail2ban, strong passwords and a non-standard port are already there.

What bugs me is detection. If one of those passwords leaks, I want to know the second someone logs in, not after files get messed with.

What are you using for instant login alerts? A PAM script that hits a Telegram or Slack webhook, or something else? And if you catch a session you don't expect, what's your move? Just `pkill -u` and `passwd -l`, or do you have something smarter hooked into PAM?
```



document fetched 2026-09-19 00:29 · canonical https://reddit.com/r/selfhosted/comments/1wjuk4k/realtime_alerts_on_ssh_logins_and_what_do_you_do
