Monitoring is the only software you buy to watch the software you already bought, and it bills by the host, so the more machines you own, the more you rent.
SigNoz is what happens when someone decides the watcher should be free too.
The best open source Datadog alternative for most teams is SigNoz: MIT-licensed outside its enterprise folder, free to self-host, and $49 a month minimum on SigNoz Cloud with no per-host and no per-user line (checked Oct 31, 2026). Grafana with Prometheus is the pick if you want the widest ecosystem, Netdata if you want per-second metrics the moment the agent installs, Zabbix if you run a large mixed fleet, OneUptime if you want the status page and the pager in one box, and VictoriaMetrics when Prometheus runs out of disk.
What Datadog actually charges
Datadog's pricing page on Oct 31, 2026 sells Infrastructure Pro at $15 per host a month billed annually ($18 on demand) and Enterprise at $23 ($27 on demand). APM is another $31 per host a month billed annually. Logs carry two meters: $0.10 per GB ingested, then $1.70 per million log events a month for 15 days of standard indexing. The free tier is five Basic hosts with one day of metric retention. On that one page the phrase “Starting At” appears 97 times. New Relic and Dynatrace meter by the gigabyte and the hour instead: the same idea in different units.
Datadog has three meters for one server: the host, the gigabytes it logs and the spans it traces. The open source stacks below have one line each, and for self-hosting it reads $0.
The open source Datadog alternatives
Of the 604 monitoring tools in the AppClap catalog (Oct 31, 2026), 69 are open source and 132 can be self-hosted. These six replace enough of Datadog to matter.
SigNoz
SigNoz is an OpenTelemetry-native stack that stores logs, metrics and traces in ClickHouse and shows them in one UI. Self-hosted SigNoz Community is free; SigNoz Cloud Teams starts at $49 a month including $49 of usage, then $0.30 per GB of logs or traces and $0.10 per million metric samples, with no host-based or user-based pricing (checked Oct 31, 2026). The licence is MIT outside the ee/ directory, with about thirty-two thousand stars (repo). It is best for a team leaving Datadog's APM and logs together. The drawback is ClickHouse: a serious database to feed and back up.
Grafana with Prometheus
Grafana is the dashboard and alerting layer, Prometheus is the metrics database it grew up with, and Loki and Tempo from the same vendor cover logs and traces. All are free to self-host. Grafana Cloud Free allows 10,000 active series and 50 GB of logs a month with 14 days of retention; Pro is a $19 a month platform fee plus usage (checked Oct 31, 2026). Grafana is AGPL-3.0 with about seventy-eight thousand stars (repo); Prometheus is Apache-2.0 with about sixty-seven thousand (repo). It is best for a team that wants the widest set of exporters and plugins. The drawback is that “Grafana” is four products, and you run all four.
Netdata
Netdata is an agent that collects per-second metrics from the host it runs on and builds its own dashboards, with anomaly detection on top. The agent is free and GPL-3.0, with about eighty-one thousand stars (repo). Netdata Cloud Community is free for up to 5 nodes; Business starts at $7 per node a month, or $15 with APM (checked Oct 31, 2026). It is best for an operator who wants to see everything on a box the minute the agent installs. The drawback is that the cloud tier is priced per node: the host meter again, at a smaller number.
Zabbix
Zabbix monitors servers, virtual machines and network devices from one server with templates, triggers and maps. The software is free and, since version 7.0, AGPLv3 (licence page); the company sells support subscriptions and Zabbix Cloud, and its pricing page returned a 404 on Oct 31, 2026. It is best for a large mixed fleet where the host count is exactly the number you refuse to pay by. The drawback is setup: hosts, templates and triggers are configured by hand, not discovered.
OneUptime
OneUptime bundles uptime checks, status pages, incidents, on-call rotation and a logs, metrics and traces backend in one product. The Free plan includes one status page, incident management and telemetry, with active monitors at $1 a month each; Growth is $22 a user a month (checked Oct 31, 2026). The licence is Apache-2.0 outside the ee/ directory, with about seven thousand eight hundred stars (repo). It is best for a small team that wants the pager and the dashboards in one install. The drawback is a far smaller community than the other five.
VictoriaMetrics
VictoriaMetrics is a Prometheus-compatible time-series database that takes less disk and memory for the same metrics. The open source version is free and Apache-2.0, with about eighteen thousand stars (repo); the vendor's pricing page returned a 404 on Oct 31, 2026. It is best as the storage under Grafana once Prometheus starts eating a server. The drawback is that it is storage, not a UI.
| Tool | Licence | Self-hosted | Paid tier (Oct 31, 2026) |
|---|---|---|---|
| SigNoz | MIT (outside ee/) | Free | Cloud from $49 a month, usage-based |
| Grafana + Prometheus | AGPL-3.0 / Apache-2.0 | Free | Cloud free tier, Pro $19 a month plus usage |
| Netdata | GPL-3.0 | Free agent | Cloud from $7 per node a month |
| Zabbix | AGPLv3 | Free | Support subscriptions, Zabbix Cloud |
| OneUptime | Apache-2.0 (outside ee/) | Free | Growth $22 a user a month |
| VictoriaMetrics | Apache-2.0 | Free | Enterprise, quote |
The wave does not stop at six: OpenObserve (AGPL-3.0, about twenty-three thousand stars) sells its cloud at $0.50 per GB with no per-host fee, and Gatus (Apache-2.0, about twelve thousand) is the smallest possible uptime page. All of them speak OpenTelemetry, which is why this is possible now: the agent is standard, so the backend can be anyone's.
Is Datadog free?
Datadog has a free tier of up to five Basic hosts with one day of metric retention, and that is the whole of it (pricing page, Oct 31, 2026). Logs, APM and a sixth host move to the paid lines: $15 per host a month for Infrastructure Pro, $31 per host for APM and $0.10 per GB for log ingestion. The free tier is a demo, not a plan.
Grafana vs Datadog: which is cheaper?
Grafana is cheaper at every size we can check, because its price does not know how many hosts you have. Grafana Cloud Free covers 10,000 series and 50 GB of logs a month at $0; Grafana Cloud Pro is $19 a month plus usage. Datadog Infrastructure Pro is $15 per host a month, so ten hosts cost $150 before a single log is ingested (all checked Oct 31, 2026). What Datadog sells for the difference is one agent, 1,000+ integrations and alerts someone else maintains.
What you actually take on
Self-hosted monitoring means you now monitor the monitor. Logs grow faster than anything else on a server, so the ClickHouse or Loki volume needs a retention policy and a backup you have tested, and upgrades of a four-part stack land in your calendar, not a vendor's.
Datadog earns its meters in three ways. Its agent covers more integrations than any of the six above, its Enterprise tier adds machine-learning alerts instead of thresholds you set by hand, and when the dashboard is down at 2am it is someone else's on-call rotation. If your host count is small, the $15 line can be cheaper than an engineer's weekend.
- 1Do you have someone who can run ClickHouse, Prometheus or a Postgres database in production?
- 2Is your host count the number that makes the Datadog bill hurt, or is it the log volume?
- 3Are you already sending OpenTelemetry, so the backend can be swapped without touching the apps?
- 4If the dashboard goes dark at 2am, is that a problem you would rather own or outsource?
We keep circling this pattern at AppClap, because the whole job is helping you find the tool that actually fits, and the catalog's open-source filter is where the per-host meter goes to die. The reaper has worked down the hallway past dictation, deploys, the CRM, the front door, the mailbox and the task board, and today he is in the server room, where the door has three meters bolted to it. Spend thirty seconds on monitor my servers and logs without paying per host, browse the Datadog alternatives side by side, and ask the question this series always ends on: do you want to rent the watcher, or own it?
Sources: the Datadog, SigNoz, Grafana Cloud, Netdata, OneUptime and OpenObserve pricing pages; the Zabbix licence page; the SigNoz, Grafana, Prometheus, Netdata, OneUptime, VictoriaMetrics, OpenObserve and Gatus repositories. All checked Oct 31, 2026.
