Datadog alternatives: 8 observability tools compared
Datadog covers almost everything and prices each product separately, mostly by host, GB or custom metric. If you want a simpler bill, or OpenTelemetry without a vendor agent, these are the tools to look at.
We make Telemetry Machine, so we're not neutral. It's listed first for that reason, and the rest are in alphabetical order. Every entry says when to choose something else.
| Tool | You pay for | Starting price | Retention included | Hosting | MCP server |
|---|---|---|---|---|---|
| Telemetry Machine (ours) | Events. Spans, log records and JSON events count as 1; 5 metric points count as 1. | $50 a month for 50M events. 14-day free trial; no free plan. | 60 days | Hosted by us | Yes, hosted, with OAuth and a Claude Code plugin |
| Better Stack | GB ingested and GB retained, plus seats for incident responders | Free: 3 GB each of logs and traces. Bundles from $45 a month in the US or $30 in Europe, with 40 GB each of traces, logs and metrics. | Free: 3 days. Bundles: 30 days for logs. | Hosted by Better Stack, or in your own bucket for $250 a month | Yes |
| ClickStack | Your own infrastructure, or ClickHouse Cloud usage | Free to self-host. Managed: 30-day trial with $300 of credit. | Up to you | Self-hosted, or managed on ClickHouse Cloud | Yes |
| Dash0 | $0.60 per million spans or logs ($0.06 to ingest, $0.54 to store) and $0.20 per million metric points ($0.02 and $0.18) | Usage only. 14-day free trial. | 30 days. Metrics: 13 months. | Hosted by Dash0 | Yes, with OAuth |
| Grafana Cloud | Active metric series, GB of logs and traces, and active users | Free: 10k series and 50 GB each of logs and traces a month. Pro: $19 a month plus usage, with logs and traces at $0.55 per GB and metrics from $6.50 per 1k series. | Free: 14 days. Pro: 30 days for logs and traces, 13 months for metrics. | Hosted, or self-host the open-source projects | Yes |
| Honeycomb | Events, with a separate allowance of metric data points | Free: 20M events a month. Pro: from $150 a month for 50M events. | 60 days. Time-series metrics: 13 months by default. | Hosted by Honeycomb | Yes, on every plan |
| New Relic | GB ingested and user seats | Free: 100 GB a month and 1 full user. Standard: $0.40 per GB beyond that, with full users at $10 for the first and $99 each after. | Free: at least 8 days. Paid plans vary by data type. | Hosted by New Relic | Yes, on all accounts |
| SigNoz | GB of logs and traces, and million metric samples | Teams: $49 a month including $49 of usage, then $0.30 per GB and $0.10 per million samples. | Logs and traces: 15 days. Metrics: 1 month. | Hosted, or self-host the open-source edition | Yes |
Telemetry Machine (ours)
Telemetry Machine brings together the best of the observability products we've loved, like the powerful investigative abilities of Honeycomb, the great UI of Datadog, and the speed of ClickHouse.
Pick it if you want to give yourself and your agents the ability to answer any question about your stack (200M spans and logs a month is $200 on Team, kept 60 days, with no per-host or per-GB charges).
Look elsewhere if: self-hosting is critical to you, or you need a free plan. We also don't have some features that bigger players have, like session replay and synthetic checks.
Better Stack
Logs, traces and metrics alongside uptime checks, status pages, error tracking and on-call.
Pick it if uptime checks, status pages and on-call matter as much to you as logs and traces.
Look elsewhere if: Anyone who handles incidents needs a responder seat, at $29 to $34 a month.
ClickStack
ClickHouse, HyperDX and an OpenTelemetry Collector, packaged as one open-source stack for logs, metrics, traces and session replays.
Pick it if you have the people to run a database and want telemetry kept in your own infrastructure.
Look elsewhere if: Self-hosting means running ClickHouse in production yourself.
Dash0
An OpenTelemetry-native platform for traces, logs and metrics, with an AI assistant called Agent0.
Pick it if you want to leave per-host pricing and stay on OpenTelemetry, and also want synthetic checks and session replay.
Look elsewhere if: Agent0 is billed separately, at $0.60 a credit.
Grafana Cloud
Hosted Prometheus-compatible metrics, Loki logs, Tempo traces and Grafana dashboards.
Pick it if you want a broad product range with a free tier, and the option of self-hosting the open-source projects later.
Look elsewhere if: Metrics are billed per active series, so high-cardinality labels cost more, and logs, metrics and traces each have their own query language.
Honeycomb
Stores traces and logs as wide events, built for high-cardinality questions, with BubbleUp to compare outliers against everything else.
Pick it if debugging high-cardinality production problems matters most, and you can live without an all-in-one platform.
Look elsewhere if: Span events and links count as events, and Pro includes only 2 SLOs.
New Relic
A long-running full-stack platform with APM, logs, metrics and traces, queried with NRQL.
Pick it if you want a full-stack platform like Datadog, billed by data and seats instead of hosts.
Look elsewhere if: Seats get expensive as more people need full access: on Pro, a full user is $349 a month.
SigNoz
An open-source, OpenTelemetry-native platform for logs, traces and metrics, built on ClickHouse.
Pick it if you want an all-in-one view that is open source and has no per-seat or per-host charge.
Look elsewhere if: Self-hosting means running ClickHouse yourself, and Teams keeps logs and traces for 15 days by default.
Telemetry Machine and Datadog, side by side
For worked costs and a row-by-row comparison, see Telemetry Machine vs Datadog.
14-day free trial. No credit card required.