S3 Service Monitoring¶

Access the S3 Service Dashboards¶

In the ARTESCA UI, open the Data Management menu and select the Data Services tab, then click Advanced Metrics.
In Grafana, click Dashboards. Open the Data Services folder and click S3 service.

Monitoring Components¶

Overview¶

Metrics	Description
Up	Number of S3 instances that are up and running.
Http requests rate	Rate of HTTP requests served.
Success rate	Percentage of successful/failing HTTP requests.
Status 200	Rate of HTTP requests served successfully.
Status 4xx	Rate of HTTP requests which failed with a client error.
Status 5xx	Rate of HTTP requests which failed with a server error.
Active Requests	Number of HTTP requests currently being served.
Injection Data Rate	Average rate of data ingestion. It corresponds to the cumulative amount of data created (>0) or deleted (<0) per second.
Injection Rate	Average rate of object ingestion. It corresponds to the cumulative number of object created (>0) or deleted (<0) per second.
OOB Injection Data Rate	Average rate of data ingested out-of-band (OOB). It corresponds to the cumulative amount of OOB data created (>0) or deleted (<0) per second.
OOB Injection Rate	Average rate of objects ingested out-of-band (OOB). It corresponds to the cumulative number of OOB objects created (>0) or deleted (<0) per second.
Buckets	Number of S3 buckets available in the cluster. This value is computed asynchronously. Its update can take up to one hour.
Objects	Number of S3 objects available in the cluster. This value is computed asynchronously. Its update can take up to one hour.
Reporter	Number of instances of Report Handler that are up and running (handles the reporting of object and bucket counts).
Last Report	Time elapsed since the last report. This delay indicates how “fresh” the Objects and Buckets values are.

Response Codes¶

Metrics	Description
Http status code over time	Number of HTTP requests grouped by return code.
Aggregated status over time	Success: Rate of successful HTTP requests. Use errors: Rate of HTTP requests that failed due to a user error (4xx). System errors: Rate of HTTP requests that failed due to a system error (5xx).

Operations¶

Metrics	Description
Request rate per S3 action	S3 request rate per action.
HTTP Method breakdown	Request distributions per HTTP method.

Latency¶

Metrics

Description

Average latencies

Average latency per requests, over time.

Overall: Average requests duration.
Upload: Average upload requests duration.
Delete: Average delete requests duration.
Download: Average download requests duration.
Multi-delete: Average multi-delete request duration.

Request time

Time to process a request for each operation.

The Y axis is the request duration. The scale uses the bucket list for that metric.

A request with a duration “x” increments all the buckets above that value. A request with a duration of 347 ms increments the buckets 400, 500, and +inf.

Latencies per S3 action

Latency per S3 operation, over time.

Data Rate¶

Metrics	Description
Bandwidth	out: Outbound bandwidth. in: Inbound bandwidth.
Avg upload chunk size by φ-quantile	Average size of uploads for each quantile. Shows if the percentage of upload requests are ≤ to the size displayed.
Avg download chunk size by φ-quantile	Average size of downloads for each quantile. Shows if the percentage of download requests are ≤ to the size displayed.