Detailed Pricing comparison of observability tools with a calculator spreadsheet
We did a cost analysis of some of the popular observability products like DataDog, New Relic, and Grafana.
Here are some key takeaways from our cost analysis:
-
User-based SaaS pricing limits the ability of engineering teams to collaborate seamlessly.
-
Custom metrics are important for having a deeper understanding of your application. If you use Datadog, your custom metrics bill can be up to 52% of your total billing. SigNoz does not charge separately for custom metrics, and with $0.1 per million samples, it is the most cost-efficient tool for metrics monitoring.
Below is the snapshot of our full stack observability cost comparison. You can have a look at our complete cost comparison analysis.
Datadog has a very complex SKU-based pricing structure. The complex billing structure makes it hard to predict how much you will be charged at the end of the month.
Recently, it was revealed that they charged a cryptocurrency company a bill of $65 million USD. A Hacker News thread discussing the report went viral, and there was an outpour of user stories around Datadog’s unpredictable billing practices.
A lot of users also pointed out how the sales team of Datadog relentlessly pursue engineers for signing up for their services.
Datadog’s pricing for custom metrics is also insane. We deep dive into it later in the blog.
Depending on the size of the engineering team, we have done cost benchmarking of three hypothetical scenarios.
- Small engineering team - 25 engineers
- Midsize engineering team - 100 engineers
- Large engineering team - 200 engineers
Datadog has a very complex SKU-based pricing structure. New Relic charges based on data ingest and user seats. Grafana charges based on the amount of telemetry data sent and user seats. SigNoz charges only on the amount of telemetry data sent.
Observability should be set up from day one. For small engineering teams, getting the most value for their money is critical. Below is a breakdown of full-stack observability cost comparison for a team of 25 engineers.
We have assumed 20 APM hosts, 50 infra hosts, and 2500 GB ingested logs.
You can find the assumptions we have taken in this sheet.
| SigNoz | Grafana | New Relic | Datadog | |
|---|---|---|---|---|
|
APM 20 APM hosts, 50 M indexed spans |
$671 | |||
|
Infra 50 infra hosts, 750k container hours, 75k custom metrics |
$5,600 | |||
|
Logs 2500 GB ingested, 1560 million log events |
$4,150 | |||
|
Logs 2500 GB ingested |
$1,000 | $1,200 | ||
|
Metrics 13 million samples per infra host (1) |
$65 | $124 | ||
|
Traces 43.8 GB per APM host |
$350 | $388 | ||
| Data Ingest | $1,178 | |||
| Users | $200 | $2,333 | ||
| Total | $1,415 | $1,912 | $3,511 | $10,421 |
| Up to 7x more value for money with SigNoz |
As your business grows, the engineering team needs to scale too. Here’s a cost comparison for a hypothetical team of 100 engineers. The tech stack consists of 125 APM hosts, 200 infra hosts, and 10,000 GB ingested logs.
| SigNoz | Grafana | New Relic | Datadog | |
|---|---|---|---|---|
|
APM 125 APM hosts, 500 M indexed spans |
$4,513 | |||
|
Infra 200 infra hosts, 1.5 M container hours, 250k custom metrics |
$17,200 | |||
|
Logs 10,000 GB ingested, 3000 million log events |
$8,500 | |||
|
Logs 10,000 GB ingested |
$4,000 | $4,950 | ||
|
Metrics 13 million samples per infra host (1) |
$260 | $494 | ||
|
Traces 43.8 GB per APM host |
$2,190 | $2,688 | ||
| Data Ingest | $5,393 | |||
| Users | $800 | $9,430 | ||
| Total | $6,450 | $8,932 | $14,823 | $30,213 |
| Up to 4.7x more value for money with SigNoz |
Large businesses need observability at scale. Here’s a cost comparison for a hypothetical team of 200 engineers. The tech stack consists of 225 APM hosts, 350 infra hosts, and 20,000 GB ingested logs.
| SigNoz | Grafana | New Relic | Datadog | |
|---|---|---|---|---|
|
APM 225 APM hosts, 2 Billion indexed spans |
$9,993 | |||
|
Infra 350 infra hosts, 2.5 M container hours, 250k custom metrics |
$45,500 | |||
|
Logs 20,000 GB ingested, 4,500 million log events |
$13,250 | |||
|
Logs 20,000 GB ingested |
$8,000 | $9,950 | ||
|
Metrics 13 million samples per infra host (1) |
$455 | $865 | ||
|
Traces 43.8 GB per APM host |
$3,942 | $4,878 | ||
| Data Ingest | $10,292 | |||
| Users | $1,600 | $18,860 | ||
| Total | $12,397 | $17,292 | $29,152 | $68,743 |
| Up to 5.5x more value for money with SigNoz |
Custom metrics give deeper insights into the performance of your application. It can help you track key application KPIs. For a robust observability setup, your engineering and DevOps teams need the flexibility and freedom to create and send as many custom metrics as needed.
But vendors like Datadog charge $0.05 per custom metric, which limits a team’s ability to send and analyze custom metrics for monitoring.
At scale, it can constitute up to 52% of your total billing.
SigNoz does not treat custom metrics any differently. The charges remain $0.1 per million samples no matter what type of metrics you send. Hence, you can create and send custom metrics with peace of mind while using SigNoz.
User-based pricing is outdated. An observability tool is used for debugging performance issues, and you never know which engineer might need it. At SigNoz, we don’t charge based on user seats.
New Relic’s user pricing can go up to $549/user. At scale, the cost of adding users can go up to 66% of the total bill.
We believe in transparent and flexible pricing. As we meet with developers, engineering leaders, and executives around the world, we realized engineering teams want two things:
- Best value for their money
- Predictability of how much they will pay
We are working tirelessly to improve our offering and value to our users. After careful examination, we identified the following issues with other tools.
| Tool | Issue |
|---|---|
| Datadog | Has the most complex pricing structure. You will never know what you might end up paying. The internet is full of many such horror stories. |
| New Relic | High user-based pricing limits collaboration. As teams become more diverse and cross-functional, you need to collaborate seamlessly. |
| Grafana | It does not have a seamless three signals (logs, metrics, traces) in a single pane experience. |