Self-hosted alternatives to Datadog

Zabbix is a free, open-source Datadog alternative for infrastructure monitoring: it watches servers, network devices, databases and websites, raises problems and alerts your team, all on your own server. Datadog charges per host and keeps your metrics on its servers; self-hosted Zabbix has no per-host fee, though it has no application tracing or log analytics like Datadog’s.

4 alternatives Free apps Our pick: Zabbix

At a glance

App Good for Code Ease of use
Zabbix IT teams and service providers watching many servers and network devices Code Open source Some setup
Beszel Small teams and IT providers watching a handful of servers and their Docker apps Code Open source Some setup
Grafana Teams that want live charts and alerts for their servers, apps or databases on one screen Code Open source Some setup
Netdata Small teams and freelancers who want a live view of one server's health without a monitoring subscription Code Partly open source Easy

Why look for an alternative to Datadog?

Datadog is a hosted monitoring and observability service: agents on your servers send metrics, logs and traces to Datadog, which shows them on dashboards and alerts your team. Infrastructure monitoring is priced per host, and logs, application tracing and other products are priced separately.

Teams look for a Datadog alternative when the bill grows with every host and product they add, or when they would rather keep monitoring data on their own servers. For watching servers, network devices, databases and websites, a self-hosted Zabbix does the job with no per-host fee, though it does not cover Datadog's tracing and log analytics. Grafana builds dashboards and alerts on top of a store you run, such as Prometheus for metrics or Loki for logs, with no per-host or per-gigabyte price. For a handful of servers and their Docker containers, Beszel is lighter: one dashboard with history charts and alerts, also with no per-host fee. Netdata gives a live, per-second view of the server it runs on, with alerts from the first minute; its agent is open source and its dashboard is closed source but free to use.

The alternatives, compared

Zabbix

Our pick

Monitor every server, switch and printer, and get alerts on problems.

Covers the infrastructure side of Datadog: agents for Linux, Windows and macOS, agentless checks over SNMP, ping, HTTP, JMX and databases, ready-made templates, automatic discovery, dashboards, trend prediction and alerts by email, Slack, Microsoft Teams, Telegram or PagerDuty. It has no application performance tracing, log analytics or real-user monitoring like Datadog's, and it takes more setup to learn.

Some setup Open source

Watch all your servers and Docker apps on one light dashboard.

Covers Datadog's infrastructure view for a handful of servers: CPU, memory, disk, network, temperature and per-container use on one dashboard, with history charts and alerts by email, Slack, Teams, Discord and more. A small agent runs on each server you watch. It has no application tracing, log search or checks from outside your servers, so it replaces the host monitoring part of Datadog, not the rest.

What you gain
  • No price per server, however many you add
  • Your servers' data stays on your own server
  • A light tool that takes little memory on each server
What is different
  • Server and container stats only, with no application tracing or log search
  • You install a small agent on each server yourself
  • Fewer integrations and dashboards than Datadog
Some setup Open source

Dashboards and alerts for your servers, apps and data.

Builds live dashboards and alerts from Prometheus, Loki, Elasticsearch, InfluxDB, PostgreSQL, MySQL, CloudWatch, Azure Monitor and more, mixed on one screen, with alerts sent by email, Slack, Teams, PagerDuty and others. Unlike Datadog it collects and stores no data itself, so you also run a store such as Prometheus for metrics or Loki for logs, and building panels means writing queries. It cannot read Datadog's data: that connector is in the paid Grafana Enterprise.

What you gain
  • No bill per server, user or gigabyte of data, only your own server
  • Your metrics and dashboards stay on your own server
  • One screen for many data sources, including your own databases
What is different
  • Grafana shows and alerts on data but stores none itself, so you also need a monitoring database or log store
  • Building charts means writing queries, though imported dashboards save most of that work
  • Connectors for Datadog, Splunk, New Relic and Snowflake, PDF reports and fine-grained permissions are in the paid Enterprise edition only
Some setup Open source

A live, self-hosted dashboard of your server's CPU, memory and disks.

Shows live, per-second charts for the server it runs on, with built-in alerts and machine-learning anomaly detection running on the server itself, and keeps history there, by default 14 days at per-second detail and longer at lower detail. The agent is open source (GPL-3.0); the dashboard is closed source but free to use. In this template it watches the server as a whole, not each app or container, and there is no APM, log search or hosted view of many servers like Datadog's without Netdata Cloud.

What you gain
  • No per-host monthly fee for watching your server
  • Per-second detail instead of averaged minutes
  • Your monitoring data stays on your server
What is different
  • This setup watches the server it runs on, not other servers, apps or containers one by one
  • Logs, top processes and AI features need a Netdata Cloud account, some of them a paid plan
  • Alert emails and chat messages need extra configuration
Easy Partly open source

Moving from Datadog

There is no import from Datadog, so you set Zabbix up afresh and run both side by side for a while. Sign in as Admin, change the password, and add your alert channels first, such as email, Slack or Microsoft Teams. Install the Zabbix agent on each server you watch and link it to the matching template, such as Linux or Windows by Zabbix agent, so its checks and alerts come ready-made. Rebuild the few dashboards your team relies on, compare alerts with Datadog for a week or two, then retire the Datadog agents.

Grafana collects no data itself, so start with where your data will live: a store such as Prometheus for metrics or Loki for logs, or a database you already have, which can sit next to Zabbix, since Grafana can also show Zabbix data through a community plugin. Add it as a data source, import a ready-made dashboard from grafana.com for the software you run, and recreate your most important Datadog monitors as Grafana alert rules.

Nothing carries over from Datadog, but Beszel takes little setup. Create your admin account, then for each server choose Add system and run the short install command Beszel gives you, the same way you installed the Datadog agent; the agent runs on Linux, FreeBSD, macOS, Windows and in Docker. Recreate the alerts you rely on, such as a server going down or a disk filling up, and set up email or a chat webhook for them. Once Beszel shows every server, remove the Datadog agents. If you also used Datadog for logs, tracing or website checks, you need a separate tool for those.

Nothing needs to move to Netdata either: it finds what to measure on its own, so there are no dashboards or agents to bring over. Open Netdata's address, sign in with the username admin and the password Cloudgate shows, and skip the offer to sign in to Netdata Cloud if you don't want an account. The built-in alerts start at once; to have them sent by email, Slack, Telegram or PagerDuty, set up notifications in Netdata's configuration.

Questions

What is the best self-hosted alternative to Datadog?

Start with Zabbix. If you want something simpler, try Netdata. All 4 are free to self-host. Zabbix, Beszel and Grafana are open source; Netdata (partly open source) keeps part of its code under a licence with conditions an open-source licence does not have.

Can Zabbix replace Datadog's APM and log management?

No. Zabbix monitors the health and metrics of servers, network devices, databases and websites. For application tracing or searching logs you need a separate tool alongside it.

Does Zabbix charge per host like Datadog?

No. Zabbix has no licence fee and no limit on hosts or metrics. Its makers sell optional support subscriptions and a hosted Zabbix Cloud, but every feature is in the free software.

Do I have to install an agent on every server?

For full server metrics, yes, the Zabbix agent goes on each server. Network devices, websites and some services can be checked without an agent, over SNMP, ping or HTTP.

Does Grafana store metrics and logs like Datadog?

No. Grafana shows and alerts on data stored elsewhere. Pair it with Prometheus for metrics, Loki for logs, or your own database.

Can Grafana show my existing Datadog data?

Not in the free version. The Datadog data source is an Enterprise plugin, so free Grafana works with open stores such as Prometheus and your own databases instead.

Can Grafana send alerts to Slack or PagerDuty?

Yes. Alert rules can notify email, Slack, Microsoft Teams, Telegram, Discord, PagerDuty, Opsgenie, webhooks and more.

Should I choose Zabbix, Grafana, Beszel or Netdata?

Choose Zabbix to monitor servers, network devices and websites with ready-made agents, templates and alerts in one app. Choose Grafana for dashboards and alerts over data you already collect in Prometheus, Loki or your own databases; it stores no data itself. Choose Beszel for a light dashboard of a handful of servers and their Docker containers, with history and alerts, that takes little setup. Choose Netdata for a live, per-second view of the one server it runs on, with alerts from the first minute and almost no setup.

Can Beszel replace Datadog APM and logs?

No. Beszel watches servers and Docker containers: CPU, memory, disk, network, temperature and their history, with alerts. It has no application tracing or log search, so keep a separate tool if you need those.

Does Beszel check my website from outside?

No, not like an uptime service. Its agents can test response time from your own servers, but for website checks from outside use an uptime monitor such as Uptime Kuma or Gatus.

How many servers can Beszel watch?

As many as you add. Beszel has no per-host fee or server limit; each server you watch runs its small agent.

Is Netdata open source?

Partly. The Netdata agent, which collects the metrics, is open source under GPL-3.0. The dashboard you use in the browser is closed source, and free to use with Netdata under the Netdata Cloud UI License.

Can Netdata watch several servers?

Not as set up here: this template watches the server it runs on as a whole. Watching more servers means installing Netdata on each one; Netdata Cloud shows them together, free for up to 5 connected servers.

Are these Datadog alternatives really free?

Yes, Zabbix, Beszel, Grafana and Netdata are free. To run them you need Cloudgate and a server. Use hosted Cloudgate, which can also provide the server, or install Cloudgate on a server of your own for free. Cloudgate's plans and prices are at cloudgate.octabyte.io/pricing. One server can run several apps.

Do I need a developer to replace Datadog?

No. Cloudgate installs the app you choose on your server and gives it a secure web address, and you use it in your browser instead of Datadog.