How much does cloud monitoring cost?

Cloud monitoring ranges from $0 (open-source self-hosted, free vendor tiers) to $200,000+ per month for enterprise deployments. The median mid-market team running 100 hosts with full observability spends $3,000 to $15,000 per month. Costs are driven primarily by host count, log volume, custom metric cardinality, and APM coverage. Verify current pricing on each vendor's pricing page before purchasing.

What is the cheapest cloud monitoring tool?

Self-hosted Prometheus and Grafana is the cheapest in pure license terms (free), but requires platform engineering capacity. Among managed services, Grafana Cloud has the most generous free tier and lowest paid-tier pricing. New Relic's 100 GB free data ingest is the next most generous. Datadog and Dynatrace are positioned at the higher end of the market.

What percentage of cloud spend goes to monitoring?

Industry research from independent and vendor sources puts the median at 7 to 8 percent of cloud infrastructure budget. Below 3 percent typically signals under-monitoring. Above 15 percent signals overspend and is a common trigger for cost-reduction projects. Microservices and Kubernetes-heavy stacks land at the higher end of the range.

Why is my Datadog bill so high?

Bill spikes usually trace to one of: custom metrics generated from high-cardinality labels (Kubernetes pod, container, namespace), log indexing charges on top of ingest charges, APM span indexing above the included 1M per host, retention upgrades beyond the 15-day default, or container counting that bills each pod as a fractional host. See vendor pricing pages and our hidden costs page for the underlying mechanics.

How can I reduce monitoring costs?

The highest-impact strategies in order: filter and sample logs at the source (saves 30 to 50 percent), control custom metric cardinality (saves 20 to 40 percent), sample APM traces at 5 to 10 percent (saves 15 to 30 percent), right-size retention to actual query patterns (saves 10 to 20 percent), and consolidate or migrate to open source where the team has capacity (saves 40 to 90 percent).

Free options

Free cloud monitoring tools 2026

Verified April 2026

Two vendor free tiers (Grafana Cloud, New Relic) can support real production observability at small scale, permanently. Self-hosted Prometheus plus Loki plus Tempo plus Grafana is free at the licence level for any scale. The OpenTelemetry stack provides vendor-neutral instrumentation. Datadog and Dynatrace free tiers are narrow trial-conversion paths.

TL;DR

Grafana Cloud and New Relic have the most generous hosted free tiers, both permanent and operationally sufficient for 10 to 30 host deployments. Self-hosted Prometheus plus Grafana plus Loki plus Tempo is free at the licence level for any scale, with engineering time as the honest cost. OpenTelemetry stack is the recommended instrumentation choice regardless of backend.

Seven free options

Free tier comparison

Each option's free-tier scope and the realistic workloads it covers. Verify against current vendor pricing pages before committing; free-tier limits change occasionally.

Option	Free tier limits	Note
Grafana Cloud	10K active series, 50 GB logs, 50 GB traces, 50 GB profiles, 3 users	Most generous hosted free tier; permanent
New Relic	100 GB unified ingest, 1 Full Platform user, unlimited Basic users	Best for APM-heavy small workloads; permanent
Datadog	5 hosts infrastructure only, 1-day retention, no APM/logs/RUM	Limited; designed to convert to paid as team grows
Splunk Cloud	500 MB/day for trial only	No permanent free tier
Dynatrace	15-day trial only	No permanent free tier
AWS CloudWatch	10 metrics, 5 GB logs, 3 dashboards, 1M API requests	Free tier per AWS account; usage above is paid
Self-hosted Prometheus + Grafana	Unlimited (engineering time excluded)	Genuinely free at licence level; cloud cost and engineering time are real

The two production-ready free tiers

Grafana Cloud and New Relic compared

Two vendor free tiers stand out as genuinely production-ready in 2026: Grafana Cloud and New Relic. Both have permanent (not trial) free tiers, both support meaningful production workloads, and both can serve as the indefinite long-term observability backend for early-stage startups and small teams without paying.

Grafana Cloud's free tier includes 10,000 active metric series, 50 gigabytes of logs (Loki), 50 gigabytes of traces (Tempo), 50 gigabytes of profiles (Pyroscope), and 3 user seats. The structure rewards Kubernetes-shaped workloads with disciplined labelling: a 10-host Kubernetes cluster with a clean Prometheus instrumentation typically produces 1,500 to 3,000 active series, comfortably inside the free tier. The 50 gigabytes of logs covers most small workloads at typical 5 to 15 GB per day. The 3-user limit is the constraint that most often forces an upgrade as the engineering team grows past three contributors.

New Relic's free tier includes 100 gigabytes of unified telemetry ingest per month and 1 Full Platform user. The single-meter ingest model means metrics, logs, traces, browser data, and custom events all count against the same 100 gigabyte budget. The structure rewards APM-heavy workloads with modest log volume; a 5-host Node.js application with full APM instrumentation and reasonable log discipline typically uses 30 to 60 gigabytes per month. Adding heavy log volume crosses the free tier quickly; teams hit the cap with surprising speed if log discipline is loose.

Both vendors compete aggressively on the free tier as a customer-acquisition strategy. The expectation is that successful customers grow past the free tier and convert to paid; the free tier is permanent precisely because conversion economics work better than trial-based conversion. For early-stage startups and small teams, the free tiers are a genuine economic gift; use them while you can.

The self-hosted free option

OpenTelemetry plus Prometheus plus Grafana

Beyond hosted vendor free tiers, the self-hosted open-source observability stack provides genuinely unlimited free observability at the licence level. The standard 2026 open-source stack is OpenTelemetry SDKs for instrumentation, OpenTelemetry Collector for routing, Prometheus for metrics, Loki for logs, Tempo for traces, Pyroscope for continuous profiling, and Grafana for dashboarding and alerting. All components are open-source under permissive licences with no licence fee at any scale.

For very small deployments (under 20 hosts, single platform engineer), the open-source stack is genuinely close to free in total cost. A single-VM deployment running Prometheus, Loki, Tempo, and Grafana on a 4-vCPU 16 GB instance costs $50 to $100 per month in cloud infrastructure. Initial setup takes 30 to 50 hours of platform engineering time; ongoing maintenance is 5 to 10 hours per month. For a one-engineer startup, the total cost of self-hosted observability at this scale is substantially less than even the free tiers of hosted vendors when measured in pure dollars; the trade-off is the engineering time and the operational responsibility for the observability stack itself.

For larger deployments, the engineering cost of self-hosted observability becomes substantial, as analysed in detail on the Datadog vs Prometheus + Grafana TCO page. At 50 to 500 hosts, the operational burden of running production-grade Prometheus + Loki + Tempo with high availability typically exceeds the hosted vendor cost at the same scale. Self-hosted is structurally cheaper at very small and very large extremes; hosted is cheaper in the middle.

The OpenTelemetry stack's value goes beyond the free-tier economics. OpenTelemetry-instrumented applications can route telemetry to any backend (CloudWatch, Datadog, New Relic, Grafana Cloud, self-hosted Prometheus, or any combination) without rewriting application code. The vendor-neutral instrumentation is the architectural insurance that makes future vendor migration feasible without a multi-month engineering programme. For new workloads, OpenTelemetry is the recommended instrumentation choice regardless of which backend the team chooses initially.

When free is not enough

Three signs you have outgrown free tiers

Free tier limits hit consistently

If you cross the Grafana Cloud 10K series limit or the New Relic 100 GB ingest limit every month, the free tier no longer fits the workload. Plan the upgrade conversation with sales; expect the paid plan to start at $200 to $500 per month for an early-paid tier.

Need for advanced features

If the team needs alert grouping with downtime exclusions, AI- assisted root-cause analysis, premium SAML SSO, audit logs, or compliance certifications (SOC 2 Type II, HIPAA, PCI DSS), free tiers typically do not include these. Time to upgrade.

Operational scale exceeds 1 user limit

New Relic's 1 Full Platform user free-tier limit is the most common upgrade trigger. Once the team has 2 plus engineers actively using observability for incident response, the per-seat paid pricing becomes unavoidable.

Cost reduction levers

Three things to maximise free-tier value

Drop labels at the agent

On Grafana Cloud, drop high-cardinality labels (pod_name, container_id, request_path) at the Grafana Agent before they reach the cloud. Stays inside the 10K series free tier for years of normal application growth.

Filter logs at the source

On New Relic, filter DEBUG and INFO logs at the application or log shipper before they enter New Relic. Keeps unified ingest below 100 GB per month indefinitely for typical small workloads.

Use OpenTelemetry instrumentation

Instrument new applications with OpenTelemetry SDKs rather than vendor-specific agents. Preserves the option to migrate between free tiers (or to self-hosted) without rewriting application code in the future.

Cross-references

/cost-for-10-hosts

Cloud monitoring cost for 10 hosts

/grafana-cloud-pricing

Grafana Cloud pricing breakdown

/new-relic-pricing

New Relic pricing breakdown

/datadog-pricing

Datadog pricing breakdown

/datadog-vs-prometheus-grafana

Datadog vs Prometheus + Grafana

/open-source-vs-paid

Open source vs paid TCO

/aws-monitoring-cost

AWS monitoring cost: native vs third-party

/calculator

Multi-vendor cost calculator

/comparison

Six-vendor comparison

Frequently asked

What is the most generous free monitoring tool in 2026?

Grafana Cloud has the most generous hosted free tier, with 10,000 active metric series, 50 GB of logs, 50 GB of traces, 50 GB of profiles, and 3 users included permanently at no cost. New Relic is a close second with 100 GB of unified telemetry ingest and 1 Full Platform user. Both vendors offer genuinely permanent free tiers (not trials) that can support real production workloads at small scale. Datadog's free tier (5 hosts of infrastructure monitoring only) is much narrower and designed to convert customers to paid plans as they scale beyond 5 hosts.

Is open-source Prometheus + Grafana actually free?

At the licence level, yes. Prometheus, Grafana, Loki, Tempo, and the broader OpenTelemetry stack are all open-source under permissive licences (Apache 2.0, AGPL-3.0 in Grafana's case). There is no licence fee at any scale. The honest costs are cloud infrastructure (typically $30 to $200 per month for a single Prometheus VM at small scale) and engineering time (30 to 50 hours of initial setup plus 5 hours per month of ongoing maintenance for a small deployment). For a one-engineer startup running 10 hosts, self-hosted Prometheus is genuinely close to free in total cost. At larger scale the engineering time becomes substantial; see the open-source-vs-paid TCO page for the honest accounting.

Can I run a real production system on a free monitoring tier?

Yes, for small workloads. A 10-host Kubernetes startup with disciplined labelling can run Grafana Cloud free tier indefinitely. A small SaaS product with under 100 GB per month of telemetry can run New Relic free tier indefinitely. The free tiers are genuinely operational, not just trials. The constraints become real as the workload grows: Grafana Cloud's 10K active series typically supports 10 to 30 hosts depending on cardinality discipline; New Relic's 100 GB ingest typically supports 5 to 20 hosts depending on log volume. Plan the transition to paid plans somewhere between 20 and 50 hosts depending on workload shape.

What are the catches with free monitoring tiers?

Three real catches. First, retention is typically shorter on free tiers (Datadog free tier is 1-day retention; Grafana Cloud free tier inherits the same retention as paid Pro at standard 13 months for metrics, more generous than peers). Second, advanced features (alerting infrastructure, AI-assisted operations, premium dashboards, enterprise security features) may be excluded or limited. Third, the support model is community-only; there is no paid support escalation for free-tier customers. For early-stage startups and side projects, these constraints rarely matter. For production systems where any of these constraints would be operationally painful, plan to pay.

What is the OpenTelemetry stack and is it free?

OpenTelemetry is the open-source observability framework that provides instrumentation libraries for major language runtimes, a collector for routing telemetry, and standardised wire formats for metrics, logs, and traces. The OpenTelemetry SDKs and the Collector are free at the licence level (Apache 2.0). Combining OpenTelemetry instrumentation with a free observability backend (Grafana Cloud free tier, New Relic free tier, or self-hosted Prometheus + Loki + Tempo) produces an end-to-end free observability stack for small workloads. The OpenTelemetry stack is the dominant open-source observability path in 2026 and the recommended instrumentation choice for new workloads regardless of backend.

Should I use multiple free monitoring tools simultaneously?

It can be a useful evaluation strategy at small scale. Deploy free-tier instrumentation on Grafana Cloud and New Relic in parallel for a quarter; evaluate operational fit, query language comfort, and dashboard ergonomics; then standardise on one as the team grows past free-tier limits. Both vendors' free tiers are generous enough to support genuine parallel evaluation without engineering compromise. Running multiple tools long-term (beyond the evaluation period) is operationally complex and not recommended; pick one and consolidate.