The number that should stop you: companies running Datadog spend, on average, around $30,000 a year on it. Not on infrastructure — on the monitoring tool for the infrastructure. Datadog's usage-based model scales with data volume, and data volume has a way of growing faster than anyone plans for.
| Free tier | Small-team monthly (typical) | Pricing model | |
|---|---|---|---|
| New Relic | 100 GB/mo ingest | often $0 pre-Series B | per-GB ingest + per-user |
| Grafana Cloud | 10k series, 50 GB logs, 50 GB traces, 3 users | ~$600–700 | usage-based |
| Datadog | 14-day trial | $500–2,000 | per-host + per-feature, usage-based |
New Relic's free tier is the most startup-friendly by a distance: 100 GB of ingest a month, free, which covers a lot of pre-Series-B teams entirely. Its per-GB model also makes costs more predictable as you scale — you're not getting billed separately for APM, logs, traces, RUM and infra the way Datadog does.
Grafana Cloud is the managed middle: you get hosted Grafana without deploying and running the open-source stack yourself, at a price well below Datadog's, with a genuinely usable free tier. The catch is cardinality — if your metrics have high-cardinality labels, the "active series" count balloons and so does the bill. Manage that and Grafana Cloud is the value play.
Datadog is the most polished and the most complete, and for a well-funded company where engineering time is more expensive than the tool, that's a fair trade. For everyone else it's a budget risk. Start on New Relic's free 100 GB tier, or self-host Prometheus + Grafana and only move to Grafana Cloud when running it yourself stops being worth the time. Reach for Datadog when someone else is paying and you've priced the annual figure honestly first.