Managed Service for OpenTelemetry uses the pay-as-you-go billing method and meters trace computation, trace storage, and metric storage separately. This topic describes the terms used in billing, the free trial, the billable items and their unit prices, billing examples, how to enable pay-as-you-go, the billing cycle, and how costs compare with self-managed open source deployments.
Notice: The prices in this topic are for reference only. For official prices, see the Pricing page.
Terms
Trace — A trace represents how one transaction or process runs through a distributed system.
Span — A span is a named, timed segment of continuous execution within a trace. Within a trace, each call is recorded as one span.
Metric — Statistical data, such as request counts, response times, and error counts of applications, APIs, and databases.
Relationship between a trace and spans — All activity that shares the same trace ID within one account counts as a single trace. A trace contains a maximum of 10 requests (spans), and each span can be up to 2 KB in size. Data beyond that size is discarded. For more information about how a trace relates to its spans, see Basic concepts of Managed Service for OpenTelemetry.
Relationship between a trace and a metric — Ingested traces are aggregated into statistical metrics. One trace generates one metric.
Billable items and unit prices
The billable items of Managed Service for OpenTelemetry are trace computation, trace storage, and metric storage. The following table lists the daily free quota and the unit price of each billable item by region.
After you enable pay-as-you-go, Managed Service for OpenTelemetry provides a daily free quota. You are not charged when your usage stays below the free quota. When your usage reaches or exceeds the free quota, you are charged for all traces and metrics.
| Billable item | Free quota | Chinese mainland | China (Hong Kong) | Regions outside China |
| Trace computation (read, write, and aggregation volume) fee | 100,000 traces/day | USD 0.13/million traces | USD 0.182/million traces | USD 0.182/million traces |
| Trace storage fee | 100,000 traces/day | USD 0.03/million traces/day | USD 0.042/million traces/day | USD 0.042/million traces/day |
| Metric storage fee | 1 million metrics/day | USD 0.0015/million metrics/day | USD 0.0021/million metrics/day | USD 0.0021/million metrics/day |
Trace computation fee
The trace computation fee depends on the number of traces that are actually ingested. The trace computation fee covers only the traffic of the current day.
Example: You ingest 1 million traces per day in a region in the Chinese mainland. The average daily fee is 1 (million traces) × 0.13 = USD 0.13.
Ingested traces are aggregated into statistical metrics, such as the latency, queries per second (QPS), and error counts of applications, APIs, and databases. To change the ingested volume, set the sampling rate on the Cluster Configurations > Sampling Storage page of the Managed Service for OpenTelemetry console.
Trace storage fee
The trace storage fee depends on the actual storage volume. The total storage volume is measured each day and depends on the data retention period: the volume measured for a day is the daily ingested volume multiplied by the data retention period. You can change the data retention period on the Cluster Configurations > Cluster Configurations page of the Managed Service for OpenTelemetry console.
Example: You ingest 1 million traces per day in a region in the Chinese mainland and store them for 15 days. The daily measured volume of stored traces is 15 (days) × 1 (million traces) = 15 million. The daily storage fee is 15 (million) × 0.03 = USD 0.45.
Metric storage fee
The metric storage fee depends on the actual storage volume. The total storage volume is measured each day and depends on the data retention period: the volume measured for a day is the daily volume of generated metrics multiplied by the data retention period. You can change the data retention period on the Cluster Configurations > Cluster Configurations page.
Example: You ingest 1 million traces per day in a region in the Chinese mainland and store them for 15 days, which generates about 1 million metrics per day. The daily measured volume of stored metrics is 15 (days) × 1 (million metrics) = 15 million. The daily storage fee is 15 (million) × 0.0015 = USD 0.0225.
Pay-as-you-go examples
The following examples use the unit prices for a region in the Chinese mainland. For a workload in another region, substitute the unit prices listed for that region in the unit price table of this topic.
Example 1: You ingest 400 million traces (about 400 million metrics) per day in a region in the Chinese mainland.
Plan 1 and Plan 2 differ only in the trace retention period. Compare their totals to see how the retention period affects the daily fee.
Plan 1: Store all trace data and metric data for 30 days.
Daily trace computation fee: 400 (million traces) × 0.13 (unit price per million traces computed) = USD 52/day
Daily trace storage fee: 400 (million traces) × 30 (days) × 0.03 (unit price per million traces stored) = USD 360/day
Daily metric storage fee: 400 (million metrics) × 30 (days) × 0.0015 (unit price per million metrics stored) = USD 18/day
Total: USD 430/day
Plan 2: Store trace data for 7 days and metric data for 30 days.
Daily trace computation fee: 400 (million traces) × 0.13 (unit price per million traces computed) = USD 52/day
Daily trace storage fee: 400 (million traces) × 7 (days) × 0.03 (unit price per million traces stored) = USD 84/day
Daily metric storage fee: 400 (million metrics) × 30 (days) × 0.0015 (unit price per million metrics stored) = USD 18/day
Total: USD 154/day
Example 2: You ingest 10 million traces (about 10 million metrics) per day in a region in the Chinese mainland.
Trace data is stored for 7 days and metric data for 30 days.
Daily trace computation fee: 10 (million traces) × 0.13 (unit price per million traces computed) = USD 1.3/day
Daily trace storage fee: 10 (million traces) × 7 (days) × 0.03 (unit price per million traces stored) = USD 2.1/day
Daily metric storage fee: 10 (million metrics) × 30 (days) × 0.0015 (unit price per million metrics) = USD 0.45/day
Total: USD 3.85/day
Enable pay-as-you-go
To enable pay-as-you-go, go to the Activate Managed Service for OpenTelemetry page.
Billing cycle
Managed Service for OpenTelemetry is billed daily and deducts the fee of the previous day at 00:00 each day.
Cost comparison with self-managed open source deployments
Most mainstream open source application performance management (APM) solutions store data for a fixed period of 7 days. Managed Service for OpenTelemetry stores data for 30 days by default.
Managed Service for OpenTelemetry prices computation and storage separately and lets you change the storage duration in real time. You can adjust the storage duration of your data as needed to control costs.
| Customer profile | Number of nodes | Daily requests | Infrastructure configuration | Self-hosting cost (USD/month) | Managed Service for OpenTelemetry(USD/month) |
| Small | 80 | 20 million requests (average complexity: 5 spans per request) | 4 Elasticsearch instances (4 vCPU, 16 GB memory, 1 TB SSD); 4 Collector instances (4 vCPU, 8 GB memory) | 864.52 | 113.29 |
| Medium | 300 | 300 million requests (average complexity: 7 spans per request) | 8 Elasticsearch instances (8 vCPU, 16 GB memory, 6 TB SSD); 8 Collector instances (4 vCPU, 8 GB memory) | 3,330.07 | 2,379.79 |
| Large | 1,000 | 1 billion requests (average complexity: 8 spans per request) | 12 Elasticsearch instances (16 vCPU, 64 GB memory, 21 TB SSD); 16 Collector instances (4 vCPU, 8 GB memory) | 9,641.81 | 9,599.15 |
The number of nodes is the number of ECS instances or Docker containers.
The number of daily requests is based on the official Apache documentation, which uses a write size of 10 KB per trace.
The infrastructure configuration is calculated with statistical data stored for 15 days and full detailed data stored for 7 days.
Managed Service for OpenTelemetry is billed by the total number of stored application requests (daily requests × retention period). It uses an open source agent, and Alibaba Cloud maintains all other components.