All Products
Search
Document Center

Object Storage Service:FAQ

Last Updated:Dec 28, 2023

This topic describes frequently asked questions about monitoring OSS resources by using CloudMonitor.

Object Storage Service (OSS) and CloudMonitor are independent Alibaba Cloud services. OSS sends data to CloudMonitor for analysis and processing. Storage and traffic usage statistics displayed in the OSS console are collected by CloudMonitor.

A delay of approximately 2 hours exists before usage data is sent to CloudMonitor. The window of data sent in a single operation cannot exceed 5 minutes. If more than 5 minutes of data is sent from OSS in a single operation, CloudMonitor rejects the data and does not allow subsequent delivery of the same data. Therefore, we recommend that you do not calculate your OSS costs based on OSS statistics that are generated by CloudMonitor. To reconcile your OSS costs, we recommend that you contact technical support.

What do I do if the status of the alert rule displays Insufficient Data?

Check data on the Monitoring Service Overview tab. If no data is generated, the Insufficient Data status occurs.

What do I do if the data in CloudMonitor shows a latency in object uploads or downloads?

Cause: Statistics in CloudMonitor are generated by sending data collection requests and do not reflect your actual environment.

Solution: Perform the following operations to troubleshoot the problem:

  1. Check whether access from the client has latency.

  2. Capture packets to analyze access data if latency exists when requesters access the corresponding bucket.

  3. Analyze access logs to check for latency.

What do I do if my own monitoring system detects latency in access to OSS?

If your own monitoring system detects that access to OSS experiences a relatively high level of latency, perform the following steps to troubleshoot the problem:

  1. Check whether your network works properly. You can ping the IP addresses of other websites to test your network connectivity.

  2. Create an Elastic Compute Service (ECS) instance that is located in the same region as the OSS bucket. Access the bucket from the ECS instance to test whether latency exists.

  3. Provide the ID of the request that experienced a high level of latency for technical support to check if latency actually exists.

  4. Capture packets to obtain upload data. Use the following parameters to analyze packets:

    tcpdump -i <Outbound port name> -s0 (Public IP address of the host and OSS endpoint) -w result.pcap

What do I do if the request validity rate is low?

Problem description: CloudMonitor displays an error message, indicating that OSS bucket p2xxx has a request validity rate of 30.51%, which is lower than 90%, and the error lasts zero minutes.

The request validity rate is the proportion of requests with HTTP status codes 2xx and 3xx to the total number of requests. You can check the number of HTTP status codes 2xx and 3xx and other HTTP status codes in the OSS console. This helps you determine whether the decrease in the request validity rate is caused by an increase in the number of requests with error response codes. You can also use real-time log query to analyze requests.

Why does CloudMonitor return HTTP status code 404?

Problem description: CloudMonitor displays a message, indicating that the number of requests for resources that do not exist in OSS bucket ***-ali of user 197*****745 is 30, and the error lasts five minutes and was fixed at 11:45.

Cause: This alert is triggered by requests that are sent to access non-existent resources and is a normal phenomenon.

Why does CloudMonitor return NoSuchWebSiteConfigration?

The 404 error is returned because the configuration related to the resource request from the client does not exist. If the configuration exists in OSS, HTTP status code 200 is returned.

Why does the OSS console display no statistics on API operations?

Statistics on API operation calls made on a specific date are available the next day. For example, statistics on API operation calls made on October 12 are available the next day (October 13).

Why is OSS usage data in CloudMonitor inconsistent with usage data in bills?

image.png

A delay of approximately 2 hours exists before usage data is sent to CloudMonitor. The window of data sent in a single operation cannot exceed 5 minutes. If more than 5 minutes of data is sent from OSS in a single operation, CloudMonitor rejects the data and does not allow subsequent delivery of the same data. We recommend that you do not use the data from CloudMonitor to reconcile your bills. We recommend that you use one of the following methods to reconcile bills:

  • Enable logging for the bucket and use logs to reconcile bills.

  • Enable real-time log query and import OSS logs for analysis.

Why is the request validity rate for a given period of time is 0 whereas the statistics on valid requests in the OSS console and the relevant logs are normal?

CloudMonitor calculates the request validity rate based on the following formula: 100% - (Number of requests with HTTP status codes 2xx + requests with HTTP status codes 3xx)/Total number of requests. We recommend that you use the OSS console or OSS logs to check whether request exceptions occur.

The cause of this problem is that OSS sends all logs of the cluster to CloudMonitor, which exceeds the window for CloudMonitor to receive data, and that CloudMonitor does not accept failed data again.