--- title: "Monitor systemd with Datadog" description: "Use Datadog's systemd integration to track the health and performance of your units." author: "Paul Gottschling" date: 2019-12-16 tags: ["infrastructure monitoring", "systemd", "processes"] blog_type_id: the-monitor locale: ko --- [Systemd](https://www.freedesktop.org/wiki/Software/systemd/) is an initialization program that [manages processes](https://thenewstack.io/unix-greatest-inspiration-behind-systemd/) on Linux systems. It was designed to [improve the performance](http://0pointer.de/blog/projects/systemd.html) of its predecessors by creating a [dependency tree](https://fedoramagazine.org/systemd-unit-dependencies-and-order/) of system components, initializing them only when needed, and using as much parallelization as possible. With systemd [becoming ubiquitous](https://www.digitalocean.com/community/tutorials/how-to-use-systemctl-to-manage-systemd-services-and-units) in Linux distributions, it's crucial that you monitor the health and performance of both systemd and the components that it manages. We're pleased to announce that our new [integration with systemd](https://docs.datadoghq.com/integrations/systemd.md) provides comprehensive visibility into system management within your Linux deployments. ![oob-v2](https://web-assets.dd-static.net/42588/1776302233-monitor-systemd-with-datadog-oob-v2.png) ## It's 10 p.m.—do you know where your units are? Systemd constructs its dependency tree by assigning system components to logically connected **units**. Some units are **services**, which [represent processes](https://manpages.debian.org/buster/systemd/systemd.service.5.en.html). Other units are **sockets**, which systemd [initializes separately](http://0pointer.de/blog/projects/socket-activation.html) from the processes that rely on them. You can use Datadog to get fine-grain visibility into the units that systemd manages and find out if they are unhealthy. Datadog's systemd integration provides detailed metrics for the [status of your units](https://www.freedesktop.org/software/systemd/man/systemctl.html#list-units%20PATTERN%E2%80%A6). You can track system-wide counts of `active`, `activating`, `inactive`, `deactivating`, and `failed` units over time (`systemd.units_by_state`), compare these counts to the total number of units, and see whether system processes have encountered issues. You can also troubleshoot process initialization by using the `unit` tag to track the `systemd.unit.active` and `systemd.unit.loaded` metrics, letting you know how long specific units have spent in each state. Custom dashboards like the one below can help you investigate possible unit-level issues. In the "Active units" graph, for example, you can see that one service initialized but, soon after, stopped being active. ![unit-status-dash](https://web-assets.dd-static.net/42588/1776302241-monitor-systemd-with-datadog-unit-status-dash.png) The integration includes an [out-of-the-box dashboard](https://app.datadoghq.com/screen/integration/30281/systemd-overview) for systemd that surveys per-unit metrics across your infrastructure, with a special focus on commonly used units like cron, SSH, and syslog. (You can customize the dashboard to show other units as well.) ## Stay on top of unit health If one of your units fails, you'll want to know as soon as possible so you can take action. Datadog's systemd integration runs a [service check](https://docs.datadoghq.com/developers/service_checks.md) that returns the state of each systemd unit and reports a `CRITICAL` status if the unit is inactive, deactivating, or failed. You can view a summary of service checks to get an overview of unit-by-unit health. ![service-checks](https://web-assets.dd-static.net/42588/1776302245-monitor-systemd-with-datadog-service-checks.png) Our integration also uses service checks to detect if Datadog can no longer connect to systemd, as well as if systemd is [unavailable](https://wiki.debian.org/systemd#systemd_hangs_on_startup_or_shutdown). You can notify your team automatically if one of our service checks reports a `CRITICAL` status by [setting alerts](https://docs.datadoghq.com/monitors/monitor_types/integration.md?tab=checkalert#integration-status). ## Keep your resources under control You can prevent a single unit from hogging system resources by [configuring systemd](http://man7.org/linux/man-pages/man5/systemd-system.conf.5.html) to impose memory and CPU limits. If a unit exceeds a resource limit, systemd will [attempt to terminate its processes](http://man7.org/linux/man-pages/man2/setrlimit.2.html). You can use Datadog to visualize per-unit resource consumption over time, helping you understand typical usage levels and set reasonable limits. ![systemd-resource-dash](https://web-assets.dd-static.net/42588/1776302253-monitor-systemd-with-datadog-systemd-resource-dash.png) ## See what runs and save what doesn't Now that Datadog integrates with systemd, you can get close-ups of health and performance on your Linux hosts, enabling you to diagnose issues with process management more easily. You can get even more visibility by using Datadog's [Live Process](https://docs.datadoghq.com/graphing/infrastructure/process.md?tab=linuxwindows) and [Live Container](https://docs.datadoghq.com/graphing/infrastructure/livecontainers.md?tab=linuxwindows) views. And for insight into the applications systemd manages, you can use any of our 1,000 integrations. To start monitoring systemd and the rest of your infrastructure, sign up for a 14-day .