Metrics

Metrics

Metrics are data on system performance. By default, many services provide free metrics for resources (e.g., Virtual Server, File Storage) which are offered as basic monitoring through ServiceWatch. Detailed monitoring can be enabled for certain resources such as Virtual Server.

Metric data is retained for 15 months (455 days), allowing access to both recent and historical data.

Namespace

Namespace is a logical separation used to distinguish and group metrics in ServiceWatch. Most Samsung Cloud Platform services use a namespace that matches the service name, which can be found in the ServiceWatch Integrated Services List.

For custom metrics, users can define a namespace that distinguishes them from other metrics within ServiceWatch, either via ServiceWatch Agent configuration or via the OpenAPI. For details on custom metrics and logs, see Custom Metrics and Logs.

Metrics

Metrics represent an ordered set of data points collected by ServiceWatch over time. Each data point consists of a timestamp, the collected value, and the unit of measurement.

For example, the CPU usage of a specific Virtual Server is one of the default monitoring metrics provided by Virtual Server. Data points can be generated by any application or activity that collects data.

By default, Samsung Cloud Platform services integrated with ServiceWatch provide free metrics for resources. Detailed monitoring for certain resources is available as a paid offering and can be enabled per service.

Metrics can only be queried in the region where they were generated. Users cannot manually delete metrics. However, if no new data is posted to ServiceWatch, metrics automatically expire after 15 months. Data points older than 15 months (455 days) are sequentially expired, and when new data points are added, those older than 15 months are deleted.

Timestamp

The timestamp of a data point indicates the time at which the data point was recorded. Each metric data point is composed of a timestamp and a value.

A timestamp consists of hour, minute, second, and date.

Metric Retention Period

ServiceWatch retains metric data as follows.

• Data points with a collection interval of 60 seconds (1 minute) are retained for up to 15 days.
• Data points with a collection interval of 300 seconds (5 minutes) are retained for up to 63 days.
• Data points with a collection interval of 3600 seconds (1 hour) are retained for up to 455 days (15 months).

Data points initially collected at a short interval are downsampled for long-term storage. For example, if data is collected at a 1‑minute interval, it is retained at 1‑minute granularity for 15 days. After 15 days, the data remains available but can only be queried at 5‑minute intervals. After 63 days, the data is further aggregated and provided at 1‑hour intervals. If metric data points need to be retained longer than the retention period, they can be separately archived using the File Download or Object Storage Export features.

Dimensions

Key-value pairs that serve as unique identifiers for metrics, allowing data points to be categorized and filtered.

For example, using the resource_id dimension of Virtual Server metrics can identify metrics for a specific server.

Collection Interval

Refers to the frequency at which data points for each service’s metrics are collected, as provided by the service-defined collection interval.

Refer to each service’s ServiceWatch metric page for the metric collection intervals.

For example, Virtual Server provides a 5‑minute collection interval for basic monitoring, and a 1‑minute interval when detailed monitoring is enabled.

Statistics

Statistics define how metric data is aggregated over a specified period. ServiceWatch provides aggregated data based on the metric data points supplied by each service to ServiceWatch. Aggregation uses the namespace, metric name, dimensions, and data point units within the defined aggregation period.

The provided statistics are Sum, Average, Minimum, and Maximum.

• Sum: The sum of all data point values collected during the period.
• Average: The sum of all data point values divided by the number of data points during the period.
• Minimum: The lowest observed value during the period.
• Maximum: The highest observed value during the period.

Units

Each statistic has a measurement unit. Examples of units include Bytes, Seconds, Count, Percent, etc.

Aggregation Period

Each statistic calculates metric data points collected over the chosen aggregation period. The aggregation period can be selected from 1 minute, 5 minutes, 15 minutes, 30 minutes, 1 hour, 3 hours, 6 hours, 12 hours, and 1 day, with a default of 5 minutes. The aggregation period is closely tied to the metric data point collection interval, and for valid aggregation results, the aggregation period must be equal to or longer than the collection interval.

For example, if the statistic is Average, the aggregation period is set to 5 minutes, and a metric with a 1‑minute collection interval is selected, data points are collected every minute and the average is computed over the 5‑minute window. Conversely, if the aggregation period is shorter than the collection interval, valid aggregation results cannot be obtained.

Downsampling is applied for long‑term retention of metric data. For example, if data is collected at a 1‑minute interval, after 15 days it can only be queried at 5‑minute granularity. Setting the aggregation period from 5 minutes to 30 minutes for such metrics may require up to 5 minutes to retrieve the downsampled data. After 63 days, the data is further aggregated and provided at 1‑hour intervals. Selecting an aggregation period from 1 hour to 1 day may require up to 1 hour for data retrieval due to aggregation processing delays.

Alerts

When creating an alert policy, a metric is evaluated over the specified evaluation window, and if it meets the condition set based on the threshold, users receive an alert notification.

Alert states are Alert, Normal, and Insufficient data.

• Alert: The metric meets the set condition.
• Normal: The metric does not meet the set condition.
• Insufficient data: No metric data exists, the metric data is missing, or data has not yet arrived.

When the alert state is Alert, if the evaluation later falls outside the condition, the alert state reverts to Normal.

For more details on alerts, see the Alert section.

Basic Monitoring and Detailed Monitoring

ServiceWatch offers two types of monitoring: basic monitoring and detailed monitoring.

Samsung Cloud Platform services integrated with ServiceWatch publish a basic set of metrics to ServiceWatch for free, providing basic monitoring. As soon as any of those services are used, basic monitoring is automatically enabled and visible in ServiceWatch.

Detailed monitoring is available for select services and incurs charges. To use detailed monitoring, it must be enabled in the service’s detailed settings.

The detailed monitoring options vary depending on the service.

• For Virtual Server, the default monitoring collection interval is 5 minutes. Enabling detailed monitoring changes the collection interval for those metrics from 5 minutes to 1 minute.
• Object Storage provides basic metrics for default monitoring, and enabling replication metrics adds additional replication metrics.

The following includes services that provide detailed monitoring and their guides.