A platform engineer deploys a new microservice. Within three days, automated alerts trigger—not because of a system crash, but because the monthly Datadog bill has suddenly spiked by 40%. This happens because a developer added a highly dynamic custom metric—such as tracking individual user IDs in a promotional campaign—without realizing how Datadog scales its pricing tiers.
This scenario is common. Datadog offers a broad feature set, but its pricing model quickly penalizes growing teams. For platform engineers watching observability spend, finding the right balance between system visibility and predictable billing is a constant struggle.
The Datadog tax: Why platform engineers seek alternatives
Datadog charges across multiple vectors—hosts, APM hosts, log ingestion, log retention, and custom metrics. This multi-dimensional billing model makes budgeting difficult.
Custom metrics are often the primary driver of unexpected costs. If you emit a metric with high cardinality—such as a user ID, IP address, or container ID—the number of unique timeseries combinations explodes. Datadog charges for these custom metrics in blocks—crossing your allocation threshold triggers steep overage fees.
Let us look at a realistic example. Suppose a team runs 20 application hosts. For illustrative purposes, let us assume they ingest 500 GB of logs monthly and track 10,000 custom metrics. Under standard baseline rates, this setup is manageable.
However, if a developer accidentally introduces a user_id tag to a custom metric during a marketing campaign with 50,000 active users, that single metric now generates 50,000 unique timeseries. The resulting overage charges for custom metrics can easily double the overall monthly bill—even though the underlying infrastructure did not grow.
Furthermore, many engineering teams pay for Datadog's entire suite—including network monitoring, database monitoring, and security posture management—when they only use basic logs and application traces. Paying for unused enterprise features creates unnecessary overhead.
When Datadog is actually worth the cost
Despite the high cost, Datadog remains a dominant player for a reason. For large enterprises with complex, multi-cloud environments, the platform provides a unified view that is difficult to replicate with smaller tools.
If your infrastructure spans AWS, Google Cloud, and on-premises bare-metal servers, Datadog acts as a single pane of glass. It offers hundreds of pre-built integrations that require almost no configuration. The platform automatically correlates metrics, traces, and logs across these different environments.
Datadog is also highly beneficial for organizations with strict enterprise compliance and security requirements. It handles complex data residency laws, role-based access control (RBAC), and audit logging out of the box.
For a massive engineering organization, the time spent configuring, hosting, and maintaining multiple open-source monitoring tools can easily exceed Datadog's subscription cost. The premium price is often justified when your primary goal is reducing engineering maintenance hours.
Open-source standards: Grafana Cloud and Prometheus
For teams that want to avoid proprietary agent lock-in, the combination of Prometheus and Grafana is the industry standard. Prometheus excels at scraping and storing time-series metrics—Grafana provides powerful visualization dashboards.
Grafana Cloud offers a managed version of this stack. It allows you to use Prometheus Query Language (PromQL) and LogQL without the operational burden of managing the underlying database clustering.
- Predictable billing: Grafana Cloud typically bills based on active series and log volume—this is often easier to predict than Datadog's host-and-metric pricing.
- Open standards: Because Grafana Cloud is built on open-source standards, you can migrate your dashboards and alerting rules back to self-hosted infrastructure if your needs change.
- Interoperability: You can query data from multiple sources—including SQL databases, Elasticsearch, and cloud APIs—and display them on a single Grafana dashboard.
This ecosystem is highly effective for teams that already use Kubernetes and want to align with CNCF (Cloud Native Computing Foundation) standards.
Developer-centric APM: Honeycomb for high-cardinality debugging
Traditional application performance monitoring (APM) tools struggle with high-cardinality data. Honeycomb takes a different approach by focusing on event-driven debugging.
Instead of aggregating metrics into pre-defined buckets, Honeycomb encourages you to send wide events. A single event can contain hundreds of fields of context—such as the customer ID, the specific query run, the container ID, and the execution time.
- High-cardinality debugging: Honeycomb allows you to query millions of unique values instantly. You can isolate a performance issue down to a single user or a specific browser version.
- Volume-based pricing: Honeycomb bills based on the total number of events ingested—it does not charge extra for custom attributes or high-cardinality tags.
- Focus on debugging over infrastructure: Honeycomb is designed for software developers who need to understand how their code behaves in production—not platform engineers who only monitor CPU and memory usage.
If your primary challenge is debugging complex distributed systems rather than monitoring server health, Honeycomb offers a more targeted solution.
Simple all-in-one alternatives: Better Stack and SigNoz
Startups and mid-sized teams often do not need enterprise-grade APM or complex multi-cloud correlation. They need simple, reliable visibility into their applications.
Better Stack combines log management, uptime monitoring, and incident response into a single platform. It uses a SQL-compatible query engine—this allows developers to search through gigabytes of logs using standard SQL queries. The setup is fast and the interface is clean—making it a strong choice for teams that want to get up and running without a steep learning curve.
SigNoz is an open-source APM tool built directly on OpenTelemetry. It provides metrics, traces, and logs in a single dashboard—mimicking much of the Datadog user experience. Because it natively supports OpenTelemetry, you do not need to install proprietary agents on your servers. You can run SigNoz on your own infrastructure or use their managed cloud offering.
These lightweight tools provide sufficient visibility for straightforward applications without the cost or complexity of an enterprise suite.
How to compare observability tools on StackMatch
Choosing the right observability stack requires balancing engineering overhead against licensing costs. Platform engineers can explore more options on StackMatch, a curated developer tools directory that offers side-by-side comparison tables and editorial reviews.
By analyzing ease of use, pricing transparency, and integration depth, you can find a tool that fits your specific infrastructure scale. Using a structured comparison helps you avoid the hidden costs of overages and proprietary agent lock-in before committing to a migration.
If you are ready to evaluate different monitoring setups, take a look at the curated observability category on StackMatch to compare features and billing structures side-by-side.
FAQs
Why does Datadog get so expensive for startups?
Datadog charges separately for different hosts, containers, logs, and custom metrics. For startups with dynamic, containerized environments, custom metrics and high-volume log ingestion can quickly trigger unexpected overage charges that scale faster than actual business growth.
Can OpenTelemetry help me avoid vendor lock-in?
Yes. By instrumenting your applications with OpenTelemetry, you decouple your telemetry collection from the backend visualization tool. This makes it much easier to test different Datadog alternatives or switch providers without rewriting your application code.
Is self-hosting Prometheus and Grafana cheaper than Datadog?
While self-hosting eliminates software licensing fees, it introduces engineering overhead. You must factor in the cost of the underlying cloud infrastructure to store metrics and the engineering hours required to maintain, scale, and secure the monitoring cluster.
What is the easiest Datadog alternative to set up?
For basic log management and uptime monitoring, Better Stack is highly regarded for its quick setup. For full-stack APM, hosted solutions like Grafana Cloud or SigNoz offer streamlined onboarding paths that do not require extensive configuration.