Monitoring & observability, in detail
No category has a wider gap between the entry price and the real bill. The billing units differ per vendor — errors, GB ingested, hosts, metric series — and each unit is a different way to be surprised. Free tiers here are genuinely usable; two smaller options join the overview five for teams optimizing for predictability.
| Tool | Free tier | Entry price | Billing unit | Watch out for | Best for |
|---|---|---|---|---|---|
| Sentry | 5k errors/mo (1 user) | Team $26/mo | Per event volume | Spans/replays are separate meters — watch the budget page | Error tracking devs set up in an afternoon |
| Grafana Cloud | 10k metric series + 50 GB logs | Pro from $19/mo + usage | Per series / GB / user | Series-count billing surprises Prometheus newcomers | Managed Prometheus/Loki without running it |
| New Relic | 100 GB ingest/mo + 1 full user | $0.40/GB beyond | Per GB + per full user | Full-platform user seats are the expensive part | Most generous free tier of any full APM |
| Datadog | 5 hosts, 1-day retention | Infra Pro $15/host/mo | Per host, per product | 100+ SKUs; the ~$65M bill story is folklore for a reason | Everything-in-one enterprise observability |
| Better Stack | 10 monitors + basic logs | Uptime ~$25/mo | Flat + per-GB logs | Newer platform; fewer integrations than the majors | Polished uptime + status pages for small teams |
| Honeybadger | Small free tier | Flat plans from ~$26/mo | Flat, by error quota | Niche vs Sentry's ecosystem | Flat-fee error tracking without usage anxiety |
| Uptime Kuma | Unlimited — OSS self-hosted | $0 | Your server | You host the thing that tells you your hosts are down | Free self-hosted uptime monitoring + status page |
Pricing models and free tiers change often — check each vendor for current terms. Compiled July 2026; the most volatile figures verified on official pricing pages.
Official pages: Sentry · Grafana Cloud · New Relic · Datadog · Better Stack · Honeybadger · Uptime Kuma
Entry price at a glance
Cheapest paid plan, USD per month — per-user plans shown for a single seat. Hatched bars are usage-based or quote-only; values marked ~ are approximate.
Notes & recent changes
- Datadog bills each product separately (infra, APM, logs, RUM, …) — the $15/host headline is rarely the real number; set ingestion controls before rollout.
- New Relic's 100 GB/month perpetual free tier survived its private-equity takeover (Nov 2023) and remains the best zero-budget APM.
- Sentry moved performance billing from transactions to spans (mid-2024) with a unified pay-as-you-go budget — cap it explicitly.
Which one should you pick?
- Start with errors only → Sentry (or Honeybadger for flat fees).
- Prometheus-shaped metrics → Grafana Cloud.
- One tool for everything and budget to match → Datadog.
- Zero budget → New Relic free tier + Uptime Kuma.