The Object Storage Service (OSS) monitoring service provides user- and bucket-level metrics for different scenarios. OSS analyzes existing metrics and collects statistics generated during a period of time, such as request status distribution by user and monthly statistics. Then, OSS generates a statistical analysis that can facilitate metric data analysis and the application of billing policies.

Note
  • Metrics such as sums, maximum values, and average values are reported at one-minute intervals. Billable usage metrics are reported at one-hour intervals.
  • CloudMonitor provides a default dashboard that displays monitoring data for Elastic Compute Service (ECS) instances. If you want to view OSS user- and bucket-level metrics, click Add View in the CloudMonitor console to add the metrics to the dashboard. For more information, see Manage monitoring charts on a custom dashboard.

User-level metrics

User-level metrics describe the overall usage and information about all OSS buckets owned by an Alibaba Cloud account. User-level metrics include monthly statistics, monitoring service overview, and request status details.
  • Monthly statistics

    The metrics for monthly statistics display the storage usage of OSS resources that you have used between 00:00:00 on the first day of the current month and the current time you view the collection. The following table describes the metrics.

    Metric name Unit Description
    Storage size Byte The total storage size used by all the buckets owned by the Alibaba Cloud account between 00:00:00 on the first day of the current month and the current time when you view the collection.
    Internet outbound traffic Byte The total amount of Internet outbound traffic sent by the user between 00:00:00 on the first day of the current month and the current time when you view the collection.
    Number of PUT requests Count The total number of PUT requests sent by the user between 00:00:00 on the first day of the current month and the current time you view the collection.
    Number of GET requests Count The total number of GET requests sent by the user between 00:00:00 on the first day of the current month and the current time when you view the collection.
  • Monitoring service overview

    Monitoring service overview shows basic service metrics. The following table describes the metrics.

    Metric name Unit Description
    Availability by User % The metric that describes the system availability of OSS. You can obtain the metric value based on the following formula: Metric value = 1 - Server error requests with the returned HTTP status code 5xx/All requests.
    Valid Request Proportion by User % The percentage of valid requests out of all requests.
    Number of Total Requests by User Count The total number of requests that are received and processed by the OSS server.
    Number of Valid Requests by User Count The total number of requests with the returned HTTP status codes 2xx and 3xx.
    Internet Outbound Traffic by User Byte The amount of outbound traffic sent over the Internet.
    Internet Inbound Traffic by User Byte The amount of inbound traffic received over the Internet.
    Intranet Outbound Traffic by User Byte The amount of outbound traffic sent over the internal network.
    Intranet Inbound Traffic by User Byte The amount of inbound traffic received over the internal network.
    CDN Outbound Traffic by User Byte The amount of outbound traffic sent over Content Delivery Network (CDN) after CDN is activated. Such outbound traffic over CDN is back-to-origin traffic.
    CDN Inbound Traffic by User Byte The amount of inbound traffic received over CDN after CDN is activated.
    Cross-region Sync Outbound Traffic by User Byte The amount of outbound traffic generated during data replication after cross-region replication (CRR) is enabled.
    Cross-region Sync Inbound Traffic by User Byte The amount of inbound traffic generated during data replication after CRR is enabled.
  • Request status details

    The metrics for request status details describe requests based on returned status codes and OSS error codes. The metrics are basic service metrics. The following table describes the metrics.

    Metric name Unit Description
    Number of Server Error Requests by User Count The total number of system-level error requests with the returned HTTP status code 5xx.
    Server Error Request Proportion by User % The percentage of requests with server-side errors out of all requests.
    Number of Network Error Requests by User Count The total number of requests with the returned HTTP status code 499.
    Network Error Request Proportion by User % The percentage of requests with network errors out of all requests.
    Number of Authorization Error Requests by User Count The total number of requests with the returned HTTP status code 403.
    Authorization Error Request Proportion by User % The percentage of requests with authorization errors out of all requests.
    Number of Resource Not Found Error Requests by User Count The total number of requests with the returned HTTP status code 404.
    Resource Not Found Error Request Proportion by User % The percentage of requests with errors indicating resources not found out of all requests.
    Number of Client Timeout Error Requests by User Count The total number of requests with the returned HTTP status code 408 or OSS error code RequestTimeout.
    Client Timeout Error Request Proportion by User % The percentage of requests with client-side timeout errors out of all requests.
    Number of Client Other Error Requests by User Count The total number of requests other than the preceding client-side error requests with the returned HTTP status code 4xx.
    Client Other Error Request Proportion by User % The percentage of requests with other client-side errors out of all requests.
    Number of Successful Requests by User Count The total number of requests with the returned HTTP status code 2xx.
    Successful Request Proportion by User % The percentage of successful requests out of all requests.
    Number of Redirected Requests by User Count The total number of requests with the returned HTTP status code 3xx.
    Redirected Request Proportion by User % The percentage of redirected requests out of all requests.

Bucket-level metrics

Bucket-level metrics include all the preceding user-level metrics and measurement and performance metrics, such as measurement reference, average latency, and successful request category.

Notice Bucket-level metrics indicate the performance of each bucket, which is different from user-level metrics. For example, the storage size metric of the user-level indicates the storage size used by all buckets owned by the Alibaba Cloud account between 00:00:00 on the first day of the current month and the current time you view the collection. The storage size metric of the bucket-level indicates the storage size used by a bucket during a specific period of time.
  • Measurement reference

    The following table describes the metrics.

    Metric name Unit Description
    Storage Size Byte The amount of storage occupied by the bucket each hour.
    Internet Outbound Traffic Byte The total amount of outbound traffic sent from the bucket over the Internet each hour.
    Number of PUT Requests Count The total number of PUT requests made to the bucket each hour.
    Number of GET Requests Count The total number of GET requests made to the bucket each hour.
  • Latency

    Request latency indicates the system performance. The metric collects only the statistics of successful requests with the returned status code 2xx. The OSS monitoring service provides minute-level average latency and maximum latency metrics to indicate the system average response capability and thrashing.

    In addition, the latency metrics are displayed based on the end-to-end (E2E) latency and server latency, which help you analyze key performance and environmental issues.

    • E2E latency refers to the end-to-end latency of successful requests made to OSS. E2E latency includes the time required for OSS to read a request, send the response, and receive acknowledgment of the response.
    • Server latency refers to the time required for OSS to process a successful request. Server latency does not include the network latency specified in E2E latency.

    The following table describes the metrics.

    Metric name Unit Description
    GetObject Request Average E2E Latency Millisecond The average E2E latency of successful GetObject requests.
    GetObject Request Average Server Latency Millisecond The average server latency of successful GetObject requests.
    GetObject Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful GetObject requests.
    GetObject Request Maximum Server Latency Millisecond The maximum server latency of successful GetObject requests.
    HeadObject Request Average E2E Latency Millisecond The average E2E latency of successful HeadObject requests.
    HeadObject Request Average Server Latency Millisecond The average server latency of successful HeadObject requests.
    HeadObject Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful HeadObject requests.
    HeadObject Request Maximum Server Latency Millisecond The maximum server latency of successful HeadObject requests.
    PutObject Request Average E2E Latency Millisecond The average E2E latency of successful PutObject requests.
    PutObject Request Average Server Latency Millisecond The average server latency of successful PutObject requests.
    PutObject Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful PutObject requests.
    PutObject Request Maximum Server Latency Millisecond The maximum server latency of successful PutObject requests.
    PostObject Request Average E2E Latency Millisecond The average end-to-end latency of successful PostObject requests.
    PostObject Request Average Server Latency Millisecond The average server latency of successful PostObject requests.
    PostObject Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful PostObject requests.
    PostObject Request Maximum Server Latency Millisecond The maximum server latency of successful PostObject requests.
    AppendObject Request Average E2E Latency Millisecond The average end-to-end latency of successful AppendObject requests.
    AppendObject Request Average Server Latency Millisecond The average server latency of successful AppendObject requests.
    AppendObject Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful AppendObject requests.
    AppendObject Request Maximum Server Latency Millisecond The maximum server latency of successful AppendObject requests.
    UploadPart Request Average E2E Latency Millisecond The average end-to-end latency of successful UploadPart requests.
    UploadPart Request Average Server Latency Millisecond The average server latency of successful UploadPart requests.
    UploadPart Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful UploadPart requests.
    UploadPart Request Maximum Server Latency Millisecond The maximum server latency of successful UploadPart requests.
    UploadPartCopy Request Average E2E Latency Millisecond The average end-to-end latency of successful UploadPartCopy requests.
    UploadPartCopy Request Average Server Latency Millisecond The average server latency of successful UploadPartCopy requests.
    UploadPartCopy Request Maximum E2E Latency Millisecond The maximum end-to-end latency of successful UploadPartCopy requests.
    UploadPartCopy Request Maximum Server Latency Millisecond The maximum server latency of successful UploadPartCopy requests.
  • Successful request category

    The metrics for successful request category indicate the capability of the system to process requests. The following table describes the metrics.

    Metric name Unit Description
    Successful GetObject Request Count The number of successful GetObject requests.
    Successful HeadObject Request Count The number of successful HeadObject requests.
    Successful PutObject Request Count The number of successful PutObject requests.
    Successful PostObject Request Count The number of successful PostObject requests.
    Successful AppendObject Request Count The number of successful AppendObject requests.
    Successful UploadPart Request Count The number of successful UploadPart requests.
    Successful UploadPartCopy Request Count The number of successful UploadPartCopy requests.
    Successful DeleteObject Request Count The number of successful DeleteObject requests.
    Successful DeleteObjects Request Count The number of successful DeleteObjects requests.
  • Mirroring-based back-to-origin

    The following table describes the metrics.

    Metric name Unit Description
    Mirror Inbound Traffic Byte The total amount of inbound traffic generated by requests with the returned value 200 and 206 for a specific origin.
    Mirror Inbound Traffic by Status Byte The amount of inbound traffic generated by requests with a specific value for a specific origin.
    Mirror Request Average Transfer Speed Byte/s The average transfer speed of requests with the returned value 200 and 206 for a specific origin.
    Mirror Request Average Transfer Speed by Status Byte/s The average transfer speed of requests with a specific value for a specific origin.
    Mirror Request Count Count The total number of requests with the returned value 200 and 206 for a specific origin.
    Mirror Request Count by Status Count The total number of requests with a specific returned value for a specific origin.
    Mirror Request Average Latency Millisecond The average latency of requests with the returned value 200 and 206 for a specific origin.
    Mirror Request Average Latency by Status Millisecond The average latency of requests with a specific returned value for a specific origin.
    Mirror Request Proportion by Status % The proportion of requests with a specific status code, such as 2xx, 3xx, 4xx, and 5xx, for a specific origin out of all requests.
    Mirror Request with a Specific Status Code Count Count The number of requests with a specific status code, such as 2xx, 3xx, 4xx, and 5xx, for a specific origin.