Monday morning, 9 AM. Lee opens his inbox and finds the monthly Datadog bill. He rubs his eyes, confirming he’s reading that number correctly. Last month the team added a few servers for load testing, and the bill more than doubled. He screenshots it and sends it to the tech chat. The CTO replies: “Look into alternatives.”
This scene has played out at countless small and medium teams over the past two years. Datadog’s product quality is undeniable, but its per-host pricing combined with separate charges for each module creates obvious cost pressure as businesses scale. When you’re using APM, log management, and infrastructure monitoring simultaneously, the bill often grows faster than the business itself.
The problem isn’t that Datadog doesn’t work well. Quite the opposite. Because it works so well, teams unconsciously enable more features, connect more data sources, and then get shocked at the end of the month. This “boiling frog” experience has driven many teams to seriously evaluate alternatives.
The Logic Behind Choosing Alternatives
Before discussing specific tools, we need to clarify one thing: what are you actually looking for?
If your core requirement is “cheap,” the answer is simple. Choose open source and self-host. But reality is usually more complex. Some teams lack the operations staff to maintain an observability infrastructure. Some teams need out-of-the-box APM capabilities. Others want to save money but won’t compromise on alerting reliability.
So the real selection dimensions are these: Does your team have the capability to operate an observability platform yourselves? Do you need full-stack observability or just logs and metrics? What is your data volume? How much migration cost can you accept?
With these questions in mind, let’s look at five alternatives worth considering right now.
SigNoz: The Open Source Full-Stack Observability Leader
SigNoz is one of the most watched projects in open source observability over the past few years. Its positioning is clear: build an open source, OpenTelemetry-based full-stack observability platform that directly competes with Datadog’s core feature set.
What does “OpenTelemetry-based” mean? Simply put, OpenTelemetry is an open standard under CNCF that defines collection specifications for traces, metrics, and logs. SigNoz natively supports this standard, meaning you won’t be locked into any vendor’s proprietary agent. Use SigNoz today, switch to something else tomorrow, and your data collection layer stays unchanged.
Back to Lee’s scenario. His team has five or six backend engineers, running over twenty microservices deployed on a self-managed Kubernetes cluster. The team has one colleague who understands operations and can handle Helm chart deployments. For a team like this, SigNoz is practically a custom fit.
SigNoz provides a unified interface for viewing traces, metrics, and logs without jumping between three different tools. Its query language is fairly intuitive, and alert configuration happens within the same platform. The overall experience aims at “one platform solving all observability problems.”
Of course, self-hosting means you handle storage scaling, version upgrades, and high availability yourself. SigNoz uses ClickHouse for storage underneath, and when data volumes are large, operating ClickHouse itself becomes work. If the team lacks experience in this area, you may need to budget some learning cost.
SigNoz also offers a cloud-hosted version priced by data volume, substantially cheaper than Datadog, suitable for teams wanting an open source solution without operating it themselves.
Grafana Cloud: The Most Mature Middle Path
If SigNoz is the “new force,” Grafana Cloud is the “established power.” The Grafana name is virtually universal in operations circles, and its dashboard capabilities remain an industry benchmark. Grafana Cloud packages Grafana, Loki (logs), Tempo (tracing), and Mimir (metrics) into a managed service.
Grafana Cloud’s free tier is quite generous: 10,000 active metric series per month, 50GB logs, and 50GB trace data. For early-stage startups or small projects, this free quota might be enough.
A real-world usage scenario looks like this: a backend team of about ten people previously used Prometheus plus self-hosted Grafana for metrics, and ELK for logs. As services multiplied, maintaining these two systems consumed substantial effort. After migrating to Grafana Cloud, they kept their familiar Grafana dashboards, switched logs to Loki’s LogQL queries, and connected tracing to Tempo. Overall operations burden lightened while they didn’t completely lose control over their data.
Grafana Cloud’s advantage lies in its mature ecosystem. Nearly every mainstream open source monitoring component can connect with Grafana, and community dashboard templates number in the thousands. If your team already uses Prometheus and Grafana, migrating to Grafana Cloud is nearly seamless.
However, Grafana Cloud’s APM capabilities are still somewhat weaker compared to Datadog. If you heavily depend on deep application performance analysis, like code-level flame graphs or database query analysis, you may need additional evaluation.
New Relic: The Veteran’s Free Tier Strategy
New Relic is a veteran in the observability space, in the same tier as Datadog as commercial products. But in recent years New Relic made a bold strategic adjustment: launching an extremely aggressive free tier.
100GB of free data ingestion per month, one free full-access user. This is quite rare among commercial observability products. New Relic’s logic is clear: get you using it first, then charge when team size expands and you need more seats.
For a small team with just two or three developers, New Relic’s free tier might be completely sufficient. You get full APM, log management, infrastructure monitoring, and synthetic monitoring, with no functional differences from the paid version. This is crucial because many tools cripple core features in free versions, but New Relic doesn’t.
New Relic’s challenge is that once teams exceed the free tier limits, its pricing model charges on both user count and data volume. For large teams, costs aren’t necessarily much lower than Datadog. So New Relic is more suitable for “small teams needing enterprise features” rather than “large teams wanting to save money.”
One more thing: New Relic has invested heavily in AI recently, adding significant intelligent alerting and anomaly detection capabilities. If your team is interested in AIOps, New Relic’s maturity in this area exceeds open source alternatives by a wide margin.
Better Stack: Modern Experience for Logs and Monitoring
Better Stack started as Logtail and Better Uptime, later merged into a unified platform. Its positioning differs slightly from the previous options, focusing more on log management and uptime monitoring rather than full-stack APM.
Better Stack’s first impression is a clean interface with modern interactions. This sounds minor, but in the observability tools space, UI experience directly impacts engineer troubleshooting efficiency. When an alert wakes you at 3 AM, opening a clean, intuitive interface versus opening a densely packed legacy console creates completely different mindsets.
Better Stack’s log search is fast, and alert configuration flows are fairly streamlined. It supports mainstream log source integrations, from Docker to Kubernetes to various cloud service logs, all connecting fairly conveniently.
Scenarios fitting Better Stack: your team primarily needs centralized log management and reliable alert notifications, without requiring deep APM distributed tracing. For example, a team running a few monolithic applications with reasonably simple business logic but needing to ensure service availability and log queryability. Better Stack provides excellent value in this scenario.
If you need complete distributed tracing and code-level performance analysis, Better Stack isn’t the most suitable choice. It’s more like one piece of the observability puzzle rather than a one-stop solution.
OpenObserve: Built for Cost-Sensitive Teams
OpenObserve is a relatively new open source project with a simple, direct core selling point: extremely low storage costs. Official claims suggest storage overhead can drop by tens of times compared to Elasticsearch, thanks to its underlying use of columnar storage and efficient compression algorithms.
For teams with especially high log volumes, storage costs are often the biggest expense. A platform generating hundreds of GB of logs daily might need a sizable Elasticsearch cluster, while OpenObserve can handle equivalent data volumes with far fewer resources.
OpenObserve supports logs, metrics, and tracing, providing a unified query interface. Deployment is relatively simple too; single-node deployment only needs one binary to run. For teams wanting quick validation, this low barrier is very attractive.
However, as a newer project, OpenObserve’s community size and plugin ecosystem are still in growth phases. If you need rich integrations and battle-tested stability, you may need to weigh this. It’s more suitable for teams with strong technical judgment, willing to try new tools, and especially sensitive to storage costs.
Core Dimension Comparison
After discussing scenarios and stories, let’s create a comprehensive comparison table. This table isn’t meant for you to “choose by table,” but to help you build a quick index in your mind, knowing each solution’s strengths and weaknesses.
| Dimension | SigNoz | Grafana Cloud | New Relic | Better Stack | OpenObserve |
|---|---|---|---|---|---|
| Open Source | ✅ Fully open | Partially (components) | ❌ Commercial | ❌ Commercial | ✅ Fully open |
| Self-Hosted | ✅ Supported | ✅ Can self-build components | ❌ SaaS only | ❌ SaaS only | ✅ Supported |
| Pricing Model | By data volume | By data + series | By users + data | By data volume | By storage |
| APM/Tracing | ✅ Complete | ✅ Tempo | ✅ Deep APM | ⚠️ Basic | ✅ Supported |
| Log Management | ✅ Built-in | ✅ Loki | ✅ Built-in | ✅ Core strength | ✅ Core strength |
| Alerting | ✅ Built-in | ✅ Mature | ✅ Intelligent | ✅ Modern | ✅ Basic |
| Free Tier | Community edition | Generous free tier | 100GB/month | Limited free | Community edition |
| Learning Curve | Medium | Low (familiar ecosystem) | Low | Low | Low |
| Team Size Fit | 5-50 people | Any size | 1-20 optimal | 3-30 people | 5-50 people |
This table reveals several patterns. First, open source solutions (SigNoz, OpenObserve) have natural advantages in cost control but require teams to possess certain operational capabilities. Second, commercial free tiers (New Relic, Grafana Cloud) are extremely friendly to small teams, offering great opportunities to “learn on someone else’s dime.” Third, no solution crushes all others across every dimension. Selection is fundamentally about making tradeoffs.
Migration Costs Cannot Be Ignored
Many teams evaluating alternatives only look at features and pricing, ignoring migration costs. This is an easy trap to fall into.
Migrating from Datadog requires considering work at several levels: agent replacement (swapping Datadog agent for OpenTelemetry Collector or the target platform’s agent), dashboard reconstruction (the dashboards you carefully built in Datadog need rebuilding on a different platform), alert rule migration (every alert’s thresholds, notification channels, and grouping logic needs reconfiguration), and the team’s adaptation period.
If you choose an OpenTelemetry-compatible solution (SigNoz, Grafana Cloud’s Tempo and Mimir both support OTLP), future tool-switching costs will be much lower. This is why OpenTelemetry is becoming the de facto industry standard. It keeps your observability data unlocked from any single platform, preserving your future freedom to switch tools.
So Which Should Your Team Choose?
Back to Lee’s story at the beginning. What did he ultimately choose? The answer is SigNoz cloud-hosted version. The reasons were practical: the team has microservice architecture requiring distributed tracing; they have technical capability but don’t want to maintain a ClickHouse cluster themselves; budget is limited but not zero; most importantly, they want to preserve future migration flexibility.
But this is his answer, not necessarily yours.
If you’re a three-to-five person startup team rapidly iterating on your product and don’t want to spend energy operating tools, New Relic’s free tier might be the most hassle-free starting point. Full features, no operations required, adequate free quota. Wait until the team grows to revisit.
If your team already uses Prometheus and Grafana and is familiar with that ecosystem, Grafana Cloud is the path of least resistance. You keep all existing knowledge and habits, just handing operational burden to a managed service.
If you’re an extremely cost-sensitive team with high log volumes but no need for deep APM, OpenObserve deserves serious evaluation. Its storage efficiency advantages become very apparent in high-volume scenarios.
If your core need is modern log and alerting experience without complex distributed tracing, Better Stack’s simplicity and efficiency will impress you.
There’s no perfect tool, only choices that fit the present moment. Most importantly, whatever you choose, try to base your data collection layer on the OpenTelemetry standard. This way, even if requirements change, the team grows, or budgets shift, you’ll have the confidence to say “switch whenever,” rather than being locked into an increasingly expensive platform with no room to move.
Stay updated with our latest AI insights
