Mail notification

Feature
1

Hello,

I can say that syncthing is stable now on 2 Synology NAS’es.

One of NAS’es has Atom CPU so syncthing is a little bit slow on this server, and opening syncthing webpage is taking very long time (about 5 minutes).

Maybe it is possible to add feature to have daily mail notification if configured mail account before?

2

What would you like the mail notification to say?

Have you tried logging the output from syncthing to a file?

3

There is a pull request which adds some caching which should make loading the UI faster.

4

I’m logging output to a file, but even with that I’d really love if syncthing could inform me even when I don’t have ssh access.

I’d like syncthing to say:

  • if other node is unavailable longer than 15 minutes => notification,
  • if puller has some problems and stop sharing => notification
  • daily / each second day / weekly report so I know that syncthing for sure is working :slight_smile:
5

A very quick way to get that right now is to write a file regularly and check it arrives the other side. The longer way would use the events interface I guess.

6

You can do this manually by bash script magic. Grepping in the logs and emailing stuff.

I don’t think this functionality will ever be part of syncthing. You can run a daemon which queries the REST API and sends emails under certain conditions.

7

I want to raise this suggestion again.

Assume we run a ST cluster on many (10+) machines, everything is nice and smooth. ST runs in the background and does it’s magic.

However sometimes issues arise (and by issues I mean every ST’s warning/error that syncthing displays, usually it is shown in the yellow/red rectangle in its WebUI).

Currently in order to become aware that there are issues with certain nodes to begin with, I need to actively express interest, and access each client and check for its statuses. And constantly doing it for large clusters is problematic, and it will never provide any sort of “near real-time” issue response.

I think that it may be very useful to be able to get notified when ST instance is having an issue.

And hopefully in a more streamlined way rather than running bash scripts on the nodes (some of them are Windows machines), and in general it’s better if ST will report its issues on its own

So mail notifications is the first idea that comes to mind (i.e. provide each node with credentials for a mail service for it to logon and send messages).

But other suggestions are welcomed. Maybe pass the errors/warnings notifications to a specified (sort of “central”) node that can be monitored explicitly.

What do you think? Is there any interest is such functionality?

8

I agree with Audrius above that it’s way out of scope… So highly unlikely, although of course there are some who are interested in it as we can see above.