S3 Bucket Monitoring

Note
This feature is introduced in VAST Cluster 5.2.1.

Metrics collected per tenant and per bucket can be queried using the VMS REST API. These metrics enable you to monitor S3 traffic, requests and errors.

Bucket level metrics are aggregated per bucket every minute and retained for 24 hours.

Tenant level performance metrics are collected every ten seconds.

Available Metrics

The following are the metrics that can be queried by external systems using the VMS REST API:

Note
Latency is measured in ms.

Object Type	Class	Metric	Description
`View`	`BucketViewMetrics`	`head_requests`	The number of HEAD requests.
		`put_requests`	The number of PUT requests.
		`get_requests`	The number of GET requests.
		`post_requests`	The number of POST requests.
		`delete_requests`	The number of DELETE requests.
		`client_error_responses`	The number of requests for which the server returned a 4xx HTTP error code.
		`server_error_responses`	The number of requests for which the server returned a 5xx HTTP error code.
Tenant	TenantMetrics	`write_iops__num_samples`	The number of samples of write I/O per second.
		`read_bw__num_samples`	The number of samples of bandwidth (uploaded/downloaded bytes per second) for read operations.
		`read_md_iops__sum_squares`	The sum of the squares of all samples of read metadata I/O operations per second.
		`write_latency__num_samples`	The number of latency samples for write operations.
		`write_bw__sum`	The sum of the sampled values of bandwidth (uploaded/downloaded bytes per second) for write operations.
		`write_latency__sum_squares`	The sum of the squares of latency samples for write operations.
		`read_md_latency__num_samples`	The number of latency samples for metadata read operations.
		`read_bw__sum_squares`	The sum of the squares of bandwidth (uploaded/downloaded bytes per second) for read operations.
		`read_latency__sum`	The sum of latency samples for read operations.
		`write_md_latency__sum`	The sum of latency samples for write metadata operations.
		`write_md_iops__num_samples`	The number of samples of metadata write I/O operations per second,
		`write_latency__sum`	The sum of latency samples for write operations.
		`write_iops__sum`	The sum of the sampled values of write I/O operations per second,
		`write_md_latency__num_samples`	The number of latency samples for metadata write operations.
		`read_md_iops__sum`	The sum of the sampled values of metadata read I/O operations per second,
		`read_bw__sum`	The sum of the sampled values of bandwidth (uploaded/downloaded bytes per second) for read operations.
		`read_md_latency__sum`	The sum of latency samples for metadata read operations.
		`write_md_iops__sum_squares`	The sum of the squares of all samples of metadata write I/O operations per second.
		`read_iops__num_samples`	The number of samples of read I/O operations per second.
		`read_iops__sum`	The sum of the sampled values of read I/O operations per second,
		`write_bw__num_samples`	The sum of the sampled values of bandwidth (uploaded/downloaded bytes per second) for write operations.
		`read_latency__num_samples`	The number of latency samples for read operations.
		`write_bw__sum_squares`	The sum of the squares of all samples of bandwidth (uploaded/downloaded bytes per second) for write operations.
		`read_iops__sum_squares`	The sum of the squares of all samples of read I/O operations per second.
		`write_iops__sum_squares`	The sum of the squares of all samples of write I/O operations per second.
		`read_latency__sum_squares`	The sum of the squares of latency samples for read operations.
		`read_md_iops__num_samples`	The number of samples of metadata read I/O operations per second.
		`write_md_iops__sum`	The sum of the sampled values of metadata write I/O operations per second,
		`write_md_latency__sum_squares`	The sum of the squares of latency samples for metadata write operations.
		`read_md_latency__sum_squares`	The sum of the squares of latency samples for metadata read operations.

How to Query the Metrics

You can use the following VMS REST API endpoints to query the metrics:

To get object IDs for tenants, send a GET request to the /tenants/ endpoint.
To get object IDs for views, send a GET request to the /views/ endpoint.
To create a custom analytics report for specified metrics, send a POST request to the /monitors/ endpoint.
To query an analytics report: GET request to the /monitors/{id}/query/ endpoint, where id is the ID of the report.
To run a query for specified metrics: GET request to the /monitors/ad_hoc_query/ endpoint.

For information about these endpoints, see the VAST Management Service (VMS) REST API Documentation. You can access this documentation at <cluster_IP>/docs.