内容简介:Simple library for notifying systemd about process state.Just add this to your
systemd
Simple library for notifying systemd about process state.
Features
NOTIFY_SOCKET journal
Installation
Just add this to your rebar.config :
{deps, [systemd]}.
Or in case of Mix project, to your mix.exs :
defp deps do
[
{:systemd, "~> 0.2"}
]
end
Then call systemd:notify(ready) when your application is ready to work/accept connections.
Non-systemd systems
This application and all functions within are safe to call even in non-systemd and non-Linux OSes. In case if there is no systemd configuration options then all functions will simply work as (almost) no-ops.
Usage
Assuming you have my_app.service unit like that
[Unit] Description=My Awesome App [Service] User=appuser Group=appgroup Type=notify # Important! This need to run in foreground and not fork. # Check documentation of your release tool. ExecStart=/path/to/my_app start WatchdogSec=10s NotifyAccess=main Restart=on-failure [Install] WantedBy=multi-user.target
(the Type=notify part is important)
You can inform systemd about state of your application. To do so just call:
systemd:notify(ready).
This will make systemctl start my_app.service to wait until application is up and running.
If you want to restart your application you can notify systemd about it with:
systemd:notify(reloading).
Message about application shutting down will be handled automatically for you.
Logs
To handle logs you have 2 possible options:
- Output data to standard error with special prefixes. This approach is much simpler and straightforward, however do not support structured logging and multiline messages.
- Use datagram socket with special communication protocol. This requires a little bit more effort to set up, but seamlessly supports structured logging and multiline messages.
This library supports both formats, and it is up to You which one (or both?) your app will decide to use.
Standard error
There is systemd_stderr_formatter which can be used with any logger that outputs to standard error, for example logger_std_h . So to use this format it is the best to define in your sys.config :
[
{kernel,
[{logger,
[{handler, default, logger_std_h, #{formatter => {systemd_stderr_formatter, #{}},
type => standard_error}}]
}}}
].
This will add required prefixes automatically for you.
Datagram socket
This one requires systemd application to be started to spawn some processes required for handling sockets, so the best way to handle it is to add predefined systemd handlers after your application starts:
logger:add_handlers(systemd).
Be aware that this one is not guaranteed to work on non-systemd systems, so if You aren't sure if that application will be ran on systemd-enabled OS then you shouldn't use it as an only logger solution in your application or you can end with no logger attached at all.
License
See LICENSE .
以上就是本文的全部内容,希望本文的内容对大家的学习或者工作能带来一定的帮助,也希望大家多多支持 码农网
猜你喜欢:本站部分资源来源于网络,本站转载出于传递更多信息之目的,版权归原作者或者来源机构所有,如转载稿涉及版权问题,请联系我们。
区块链核心算法解析
【瑞士】Roger Wattenhofer(罗格.瓦唐霍费尔) / 陈晋川、薛云志、林强、祝庆 / 电子工业出版社 / 2017-8 / 59.00
《区块链核心算法解析》介绍了构建容错的分布式系统所需的基础技术,以及一系列允许容错的协议和算法,并且讨论一些实现了这些技术的实际系统。 《区块链核心算法解析》中的主要概念将独立成章。每一章都以一个小故事开始,从而引出该章节的内容。算法、协议和定义都将以形式化的方式描述,以便于读者理解如何实现。部分结论会在定理中予以证明,这样读者就可以明白为什么这些概念或算法是正确的,并且理解它们可以确保实现......一起来看看 《区块链核心算法解析》 这本书的介绍吧!