Datadog's multi-dimensional pricing catches teams off guard — custom metrics alone can hit 52% of your bill. Here are five alternatives that trade lock-in for predictable costs, open-source flexibility, or both.
Datadog is excellent at what it does. It's also one of the most expensive line items on a modern engineering budget. Analysis of 1.7 million SaaS purchases found the average Datadog customer spends $30,809 per year, and observability overall accounts for 16% of total SaaS spend — the second-largest category after sales and marketing.3
The core problem isn't the sticker price. It's the multi-dimensional billing model. Datadog charges separately for infrastructure monitoring ($15–23/host/month), APM (an additional $31/host/month), log indexing ($1.70/million events), and custom metrics that exceed your per-host allotment.24 At scale, custom metrics alone can account for up to 52% of the total bill.2 A team running 50 services with 2TB of monthly logs can expect $8,000–15,000/month on Datadog, while the same workload on Grafana Cloud runs $1,500–4,500.5
Three pain points push teams toward alternatives:
Here are five alternatives that address these pain points — from fully open-source to commercial SaaS — compared on what actually matters: pricing model, signal coverage, deployment flexibility, and OTel support.
(Disclosure: AskBuy earns affiliate commissions on some of the tools below. That doesn't change our recommendations — we picked these based on the research, not the payout.)
SigNoz was built explicitly as a Datadog alternative, and it shows. It's OpenTelemetry-native from the ground up, which means you instrument once with standard OTel SDKs and get metrics, traces, and logs in a single unified view — no proprietary agents.4
The self-hosted version is free. The cloud offering charges $0.30/GB for logs and traces, a fraction of Datadog's per-event indexing costs.4 For teams already investing in OTel instrumentation, SigNoz is the most direct swap: same signals, simpler billing, no lock-in.
Where it shines: Teams who want one tool for all three signals, prefer open-source, and are comfortable with (or want) self-hosting.
Where it falls short: Smaller community than Grafana's ecosystem. Fewer pre-built integrations than Datadog's 800+ out-of-the-box checks.
The Grafana stack — Prometheus for metrics, Loki for logs, Tempo for traces — is the most widely adopted open-source observability combo.2 Grafana's dashboarding is the gold standard, and the real product here is portability: you can self-host the whole stack, run it on Grafana Cloud ($228/year + usage), or mix and match.13
Real-world cost data puts Grafana Cloud at $5K–20K/year for mid-market teams, compared to $30K+ for Datadog at similar scale.35 For a startup with 10 services and 200GB of monthly logs, Grafana Cloud runs $240–440/month versus Datadog's $800–1,400.5
Where it shines: Teams who want best-in-class visualization, the flexibility to self-host or use cloud, and an escape hatch at every layer.
Where it falls short: It's three separate components, not one product. Operational overhead is higher than a unified platform, especially self-hosted.
New Relic flipped the pricing model: instead of per-host and per-event charges, they bill per user plus data ingestion. The result is far more predictable for small and mid-size teams.1 They also offer a generous 100GB free tier, which covers a lot of ground for startups.5
Mid-market spend runs $15K–40K/year — still cheaper than Datadog's average, with a billing model that's easier to reason about.3 New Relic's query language (NRQL) is developer-friendly, and the APM experience is polished out of the box.1
Where it shines: Teams who want a commercial SaaS with predictable per-user pricing and don't want to manage infrastructure.
Where it falls short: Still SaaS-only — no self-hosting option for data sovereignty needs. Usage-based data ingestion can still surprise you at high volume.
SkyWalking is fully Apache 2.0 licensed with zero enterprise paywall. There's no "open-core" upsell, no feature-gated tier — every capability is free at any scale.4
It excels at APM for polyglot microservices, with strong service topology visualization that maps dependencies and latency across your service mesh. It supports metrics, traces, and logs, and integrates with OTel collectors.4
Where it shines: Teams running microservices who want APM with service topology maps and have zero budget for observability tooling.
Where it falls short: Less polished than commercial alternatives. The UI and query experience lag behind Datadog or New Relic. Community is active but smaller than Grafana's.
If your infrastructure lives entirely in AWS, CloudWatch is already there. Logs, metrics, and traces (via X-Ray) are native to the platform, with no additional agents to install for many AWS services.1
Pricing is usage-based and shows up on your existing AWS bill — no separate vendor relationship to manage. For teams deep in the AWS ecosystem, the integration depth is hard to beat.1
Where it shines: AWS-native teams who want zero-friction setup and consolidated billing.
Where it falls short: AWS-only. Cross-cloud or hybrid teams will find it limiting. The query experience (CloudWatch Logs Insights) is basic compared to dedicated observability platforms. Costs can still spiral with high log volume.
| Pricing model | Signal coverage | Deployment | OTel support | |
|---|---|---|---|---|
| SigNoz | $0.30/GB or free self-host | Metrics, traces, logs | Self-host or cloud | Native |
| Grafana Loki | Free self-host / $228/yr+ cloud | Logs, metrics, traces | Self-host or cloud | Full |
| New Relic | Per-user + usage, 100GB free | Metrics, traces, logs | SaaS only | Full |
| Apache SkyWalking | Free (Apache 2.0) | Metrics, traces, logs | Self-host | Partial |
| AWS CloudWatch | Usage-based on AWS bill | Logs, metrics, traces | AWS-native SaaS | Partial |
The key trade-off: open-source stacks (SigNoz, Grafana, SkyWalking) give you cost control and data sovereignty but require operational effort. Commercial alternatives (New Relic, CloudWatch) reduce operational overhead but keep you in a SaaS relationship with its own lock-in dynamics.
For most teams, the recommended path in 2026 is: instrument with OpenTelemetry, then choose your backend based on budget and sovereignty needs.5 OTel makes the backend swappable — which is exactly the leverage Datadog's pricing model was designed to take away.
This page was written by the engine and the engine is still on the line. The conversation below picks up where the article stops.
Yes — the picks above are the engine's current verdicts. Ask a sharper version of this question below and you'll get a custom answer with the latest pricing.