Monitoring indicators reference

Last Updated: Oct 10, 2017

OSS indicators can be monitored at the user level or the bucket level based on application scenarios.

In addition to common chronological metric indicators, the system analyzes and collects statistics on the existing metric indicators for easy user observation of metric data and matching of billing policy. Statistical indicators over a specified period of time are provided, such as request status distribution and metering statistics of the month. This reference guide describes the indicators in detail.

All indicators are collected at the minute-level (per minute) except for metering and statistical indicators. Metering indicators are collected at the hour-level (per hour).

User-level indicators

User-level indicators consist of three monitoring indicator details: current-month metering statistics, service monitoring overview, and request state details.

Current-month metering statistics

Metering statistics of the current month are collected from 00:00 on the first day of the month to the metering cutoff time as indicated in the same month.

Details of the metering indicators currently available are as follows:

Indicator Unit Description
Storage size Byte Size of the total storage occupied by all buckets of a specified user before the metering statistic collection deadline
Internet outbound traffic Byte Total Internet outbound traffic of the user from 00:00 of the first day of the current month to the metering statistic collection deadline.
Put requests Times Total number of Put requests of the user from 00:00 of the first day of the current month to the metering statistic collection deadline.
Get requests Times Total number of Get requests of the user from 00:00 of the first day of the current month to the metering statistic collection deadline.

Service monitoring overview

Indicators in service monitoring overview are basic service indicators. Details of service monitoring overview indicators are as follows:

Indicator Unit Description
Availability % An indicator showing the system availability of using the storage service. It is obtained through the equation: Availability = 1 - percentage of requests with server-end errors (indicated by a return code 5xx) in all requests.
Valid requests rate % Percentage of valid requests in all requests. For details about valid requests, see the following description.
Requests Times Total number of requests received and processed by the OSS server
Valid requests Times Total number of requests whose return code is 2xx or 3xx.
Internet outbound traffic Byte Downstream Internet traffic
Internet inbound traffic Byte Upstream Internet traffic
Intranet outbound traffic Byte Downstream intranet traffic of the service system
Intranet inbound traffic Byte Upstream intranet traffic of the service system
CDN outbound traffic Byte Downstream CDN traffic when CDN acceleration service is activated, that is, the back-to-source traffic
CDN inbound traffic Byte Upstream CDN traffic when CDN acceleration service is activated
Outbound traffic of cross-region replication Byte Downstream traffic generated in the data replication process when the cross-region replication function is activated
Inbound traffic of cross-region replication Byte Upstream traffic generated in the data replication process when the cross-region replication function is activated

Request state details

Request state details indicators are requested monitoring information based on the return status code, or OSS error code, associated with the different requests. Details of request state details indicators are as follows:

Indicator Unit Description
Server-site error requests Times Total number of requests with system-level errors indicated by a return code 5xx
Server-site error requests rate % Percentage of requests with server-end errors in all requests
Network error requests Times Total number of requests whose HTTP status code is 499
Network error requests rate % Percentage of requests with network errors in all requests
Client-end authorization error requests Times Total number of requests with a return code 403
Client-end authorization error requests rate % Percentage of requests with client-end authorization errors in all requests
Client-end error requests indicating resource not found Times Total number of requests with a return code 404
Client-end error requests rate indicating resource not found % Percentage of requests with client-end errors indicating resource not found in all requests
Client-end timeout error requests Times Total number of requests whose return status code is 408 or return OSS error code is RequestTimeout
Client-end timeout error requests rate % Percentage of requests with client-end timeout errors in all requests
Other client-end error requests Times Total number of requests with other client-end errors indicated by a return code 4xx
Other client-end error requests rate % Percentage of requests with other client-end errors in all requests
Successful requests Times Total number of requests whose return code is 2xx.
Successful requests rate % Percentage of successful requests in all requests
Redirect requests Times Total number of requests whose return code is 3xx.
Redirect requests rate % Percentage of redirect requests in all requests

Bucket-level indicators

Bucket-level indicators are used to monitor OSS operations of specific buckets and are applicable for business scenarios. As well as current-month metering statistics and basic service indicator items such as service monitoring overview and request state details (which can be monitored at the account level) bucket-level indicators include metering indicators and performance indicators such as metering reference, latency, and successful request operation categories.

Service monitoring overview

Similar to the user-level description, the service monitoring overview indicators are basic indicators, but use metric data that is displayed at the bucket-level.

Request state details

Similar to the user-level description, the request state details indicators use metric data that is displayed at the bucket-level.

Current-month metering statistics

Statistical methods are similar to those listed in current-month metering statistics at the user level, but the former collects resource usage statistics at the bucket level. Details of current-month metering statistics at the bucket-level are as follows:

Indicator Unit Description
Storage size Byte Size of storage occupied by a specified bucket before the metering statistic collection deadline
Internet outbound traffic Byte Total Internet outbound traffic of a specified bucket from 00:00 of the first day of the current month to the metering statistic collection deadline.
Put requests Times Total number of Put requests of a specified bucket from 00:00 of the first day of the current month to the metering statistic collection deadline.
Get requests Times Total number of Get requests of a specified bucket from 00:00 of the first day of the current month to the metering statistic collection deadline.

Metering indicators

Metering indicators are monitored chronologically, and are collected and aggregated at the hour-level. Details of metering indicators are as follows:

Indicator Unit Description
Storage size Byte Average size of storage used by a specified bucket in an hour
Internet outbound traffic Byte Total Internet outbound traffic of a specified bucket in an hour.
Put requests Times Total number of Put requests of a specified bucket in an hour.
Gut requests Times Total number of Gut requests of a specified bucket in an hour.

Latency

Latency Request latency directly reflects the system performance and is monitored using two indicators: average latency and maximum latency. The indicators are collected and aggregated at the minute-level.

Moreover, indicators can be classified based on the OSS API request operation type to more specifically reflect the performance of the system responding to different operations. Only APIs involving data operations in bucket-related operations (excluding meta operations) are monitored currently.

Besides, in order to facilitate analyzing performance hotspots and environmental problems, latency monitoring indicators are collected from two different links of E2E and the server, in which:

  • E2E latency refers to the E2E latency of sending a successful request to OSS, including the processing time OSS requires to read the request, send a response, and receive a response confirmation message.
  • Server latency is the latency of OSS processing a successful request, excluding the network delay involved in E2E latency.

Note that performance indicators are used to monitor successful requests (with a return status code 2xx).

The following table lists specific metric indicator items:

Indicator Unit Description
Average E2E latency of GetObject requests Millisecond Average E2E latency of successful requests whose request API is GetObject
Average server latency of GetObject requests Millisecond Average server latency of successful requests whose request API is GetObject
Maximum E2E latency of GetObject requests Millisecond Maximum E2E latency of successful requests whose request API is GetObject
Maximum server latency of GetObject requests Millisecond Maximum server latency of successful requests whose request API is GetObject
Average E2E latency of HeadObject requests Millisecond Average E2E latency of successful requests whose request API is HeadObject
Average server latency of HeadObject requests Millisecond Average server latency of successful requests whose request API is HeadObject
Maximum E2E latency of HeadObject requests Millisecond Maximum E2E latency of successful requests whose request API is HeadObject
Maximum server latency of HeadObject requests Millisecond Maximum server latency of successful requests whose request API is HeadObject
Average E2E latency of PutObject requests Millisecond Average E2E latency of successful requests whose request API is PutObject
Average server latency of PutObject requests Millisecond Average server latency of successful requests whose request API is PutObject
Maximum E2E latency of PutObject requests Millisecond Maximum E2E latency of successful requests whose request API is PutObject
Maximum server latency of PutObject requests Millisecond Maximum server latency of successful requests whose request API is PutObject
Average E2E latency of PostObject requests Millisecond Average E2E latency of successful requests whose request API is PostObject
Average server latency of PostObject requests Millisecond Average server latency of successful requests whose request API is PostObject
Maximum E2E latency of PostObject requests Millisecond Maximum E2E latency of successful requests whose request API is PostObject
Maximum server latency of PostObject requests Millisecond Maximum server latency of successful requests whose request API is PostObject
Average E2E latency of AppendObject requests Millisecond Average E2E latency of successful requests whose request API is AppendObject
Average server latency of AppendObject requests Millisecond Average server latency of successful requests whose request API is AppendObject
Maximum E2E latency of AppendObject requests Millisecond Maximum E2E latency of successful requests whose request API is AppendObject
Maximum server latency of AppendObject requests Millisecond Maximum server latency of successful requests whose request API is AppendObject
Average E2E latency of UploadPart requests Millisecond Average E2E latency of successful requests whose request API is UploadPart
Average server latency of UploadPart requests Millisecond Average server latency of successful requests whose request API is UploadPart
Maximum E2E latency of UploadPart requests Millisecond Maximum E2E latency of successful requests whose request API is UploadPart
Maximum server latency of UploadPart requests Millisecond Maximum server latency of successful requests whose request API is UploadPart
Average E2E latency of UploadPartCopy requests Millisecond Average E2E latency of successful requests whose request API is UploadPartCopy
Average server latency of UploadPartCopy requests Millisecond Average server latency of successful requests whose request API is UploadPartCopy
Maximum E2E latency of UploadPartCopy requests Millisecond Maximum E2E latency of successful requests whose request API is UploadPartCopy
Maximum server latency of UploadPartCopy requests Millisecond Maximum server latency of successful requests whose request API is UploadPartCopy

Successful request operation categories

In conjunction with latency monitoring, the monitoring of successful requests reflects the system capability of processing access requests to a certain extent. Similarly, only APIs involving data operations in bucket-related operations are monitored currently.

The following lists specific indicator items:

Indicator Unit Description
Successful GetObject requests Times Number of successful requests whose request API is GetObject
Successful HeadObject requests Times Number of successful requests whose request API is HeadObject
Successful PutObject requests Times Number of successful requests whose request API is PutObject
Successful PostObject requests Times Number of successful requests whose request API is PostObject
Successful AppendObject requests Times Number of successful requests whose request API is AppendObject
Successful UploadPart requests Times Number of successful requests whose request API is UploadPart
Successful UploadPartCopy requests Times Number of successful requests whose request API is UploadPartCopy
Successful DeleteObject requests Times Number of successful requests whose request API is DeleteObject
Successful DeleteObjects requests Times Number of successful requests whose request API is DeleteObjects
Thank you! We've received your feedback.