Bugs are evident™ - MATRIX️
v22.1.0
Get a notification via e-mail when a SystemD service becomes failed. If you enable boot, every times to startup the notifier, it sends failed and not-found (trigger) services so you can remove them from SystemD and get cleaned.
Also it is easy to configure additional triggers like running, stopped or all the whole SystemD changes via LoadState / other properties and trigger a result / status and notify or add in different notifier like Twitter / Facebook etc.
It is easy to evolve the functions. I just created for my server to get failed services via notify e-mail, but if you need additional functions, please fork and pull. It is easy to add in anything or change services etc... All DBus based and async/await wrappers.
It is a Linux/Unix/BSD (tested only in Debian/Testing repo) based SystemD manager. Notifies via e-mail with NodeMailer, it polls via an interval as a watchdog. It also has a wrapper for DBus to manage services and via events as well if you do not like polling. I guess watchdog will be replaced 100%.
#you probably might need a c++11 if it is old,
#for additional requirements check out .travis.yml
sudo apt-get install libdbus-1-dev libglib2.0-dev
Please do not use yarn
, because it asks for sudo
and prompt (unless you are root
).
npm install p3x-systemd-manager --save
#!/usr/bin/env node
const systemd = require('p3x-systemd-manager');
const settings = systemd.lib.getSettings();
if (settings === false) {
return;
}
systemd.boot(settings);
systemd.notifier(settings);
This notifies changes in the SystemD via e-mail. Right now it polls, so that it gets all changes. It task about 30-50 milliseconds per run on my 3.3 GHz Pentium 2 cores, not too much. All automatic, requires email and a few tweaks as you want.
const Watchdog = require('p3x-systemd-manager').watchdog;
const settings = require('./settings.json');
const watchdog = Watchdog(settings);
watchdog.run();
Feb 22 11:41:31 server systemd[1]: Started p3x-watchdog.
Feb 22 11:41:32 server watchdog[2196]: started
Feb 22 11:41:32 server watchdog[2196]: watchdog type(s): service
Feb 22 11:41:32 server watchdog[2196]: ping: 2 hours
Feb 22 11:41:32 server watchdog[2196]: interval: 10 seconds
Feb 22 11:41:32 server watchdog[2196]: command: systemctl --plain --no-pager --no-legend --type=service
Feb 22 11:41:32 server watchdog[2196]: ping - 51 items - every 2 hours
Feb 22 11:41:32 server watchdog[2196]: Mail is working.
git clone https://github.com/patrikx3/systemd-manager.git
cd systemd-manager
sudo apt-get install libdbus-1-dev libglib2.0-dev
npm install
./notifier settings.json
#it is used to be a watchdog, polling
./watchdog settings.json
Checkout artifacts/setttings.json
filter.type
: Array, can be empty, actual man systemctl
type. Service is the safest. Not always working when
you fine tune, some are weird.
nodemailer.config
: Exact nodemailer config, any of that.
interval, ping
: Uses npm milliseconds
framework for turn into actual milliseconds from a string. This is
for Watchdog
, not needed anymore.
sudo
: for the watchdog either you need to use root, or via sudo (true|false
). For SystemD needs root, but
you can use another user, and it will use sudo then when polling.
For SystemD DBus notifier you need to use root anyway. I think it cannot do anything else so it's safe to take over the system, also it's internal, no web interface for now.
{
"debug": false,
"filter": {
"type": [
"service"
],
"exclude": [],
"include": [],
"trigger": {
"SubState": [
"failed"
]
}
},
"boot": {
"enabled": true,
"trigger": {
"SubState": [
"failed"
]
}
},
"moment": "LLL",
"prefix": "P3X-SYSTEMD-NOTIFIER",
"dbus": {
"address": "unix:path=/run/dbus/system_bus_socket",
"display": ":0"
},
"interval": "watchdog only",
"interval": "10 seconds",
"ping": "watchdog only",
"ping": "2 hours",
"sudo": "watchdog only",
"sudo": false,
"email": {
"to": "system@localhost",
"from": "system@localhost"
},
"nodemailer": {
"singleton": true,
"config": {
"host": "mail.localhost",
"port": 465,
"secure": true,
"auth": {
"user": "system@localhost",
"pass": "unknown"
}
}
},
"ignoreErrors": [
"The maximum number of pending replies per connection has been reached",
"No introspectable"
]
}
There are 2 errors that I could not catch or why it happens, so I created a hack:
const settings = {
"ignoreErrors": [
"The maximum number of pending replies per connection has been reached",
"No introspectable"
]
}
process.on("unhandledRejection", function (err) {
if (err && err.message && settings.ignoreErrors.includes(err.message)) {
console.warn('ignoring error', err)
} else {
// user your own logic
console.error(err)
process.exit(-1)
}
});
The two error messages:
"The maximum number of pending replies per connection has been reached"
"No introspectable"
If you appreciate our work, consider starring this repository or making a donation to support server maintenance and ongoing development. Your support means the world to us—thank you!
Our server may occasionally be down, but please be patient. Typically, it will be back online within 15-30 minutes. We appreciate your understanding.
All my domains, including patrikx3.com and corifeus.com, are developed in my spare time. While you may encounter minor errors, the sites are generally stable and fully functional.
Version Structure: We follow a Major.Minor.Patch versioning scheme:
- Major: Corresponds to the current year.
- Minor: Set as 4 for releases from January to June, and 10 for July to December.
- Patch: Incremental, updated with each build.
Important Changes: Any breaking changes are prominently noted in the readme to keep you informed.
P3X-SYSTEMD-MANAGER Build v2024.4.121