Skip to content

Listing Metrics with Utapi

Jump to bottom
Bennett Buchanan edited this page Jan 12, 2017 · 14 revisions

To make a successful request to Utapi you would need

  1. IAM user with a policy giving access to Utapi
  2. Sign request with Auth V4

IAM user with a policy giving access to Utapi

Note: The examples here use AWS CLI but any AWS SDK is capable of these actions.

endpoint-url: This would be https://<host>:<port> where your Identity(IAM) Server is running.

  1. Create an IAM user

    aws iam --endpoint-url <endpoint> create-user --user-name utapiuser

  2. Create access key for the user

    aws iam --endpoint-url <endpoint> create-access-key --user-name utapiuser

  3. Create a managed IAM policy

    sample utapi policy

    cat - > utapipolicy.json <<EOF
{
	"Version": "2012-10-17",
	"Statement": [
		{
			"Sid": "utapiMetrics",
			"Action": [ "utapi:ListMetrics" ],
			"Effect": "Allow",
			"Resource": "arn:scality:utapi:::buckets/*"
		}
	]
}
EOF

    The user can also be restricted to only some buckets (foo bucket for example) by changing the Resource property to "Resource": "arn:scality:utapi:::buckets/foo"

    aws iam --endpoint-url <endpoint> create-policy --policy-name utapipolicy \
 --policy-document file://utapipolicy.json

    A sample output of the above command would look like

    {
    "Policy": {
        "PolicyName": "utapipolicy",
        "CreateDate": "2017-06-01T19:31:18.620Z",
        "AttachmentCount": 0,
        "IsAttachable": true,
        "PolicyId": "ZXR6A36LTYANPAI7NJ5UV",
        "DefaultVersionId": "v1",
        "Path": "/",
        "Arn": "arn:aws:iam::0123456789012:policy/utapipolicy",
        "UpdateDate": "2017-06-01T19:31:18.620Z"
    }
}

    The arn property of the response, which we call , will be used in the next step to attach the policy to the user.

  4. Attach user to the managed policy

    aws --endpoint-url <endpoint> iam  attach-user-policy --user-name utapiuser --policy-arn <policy arn>

Now the user utapiuser has access to ListMetrics request in Utapi on all buckets.

Signing request with Auth V4

There are two options here.

You can generate V4 signature using AWS SDKs. See the following urls for reference.

http://docs.aws.amazon.com/general/latest/gr/sigv4_signing.html http://docs.aws.amazon.com/general/latest/gr/sigv4-signed-request-examples.html

or

You can use a nifty command line tool available in Scality's S3.

You can git clone S3 repo from here https://github.com/scality/S3.git and follow the instructions in README to install the dependencies.

If you have S3 running inside a docker container you can docker exec into the S3 container as

docker exec -it <container id> bash

and then run the command

node bin/list_bucket_metrics

It will generate the following output listing available options.

Usage: list_bucket_metrics [options]

  Options:

    -h, --help                    output usage information
    -V, --version                 output the version number
    -a, --access-key <accessKey>  Access key id
    -k, --secret-key <secretKey>  Secret access key
    -b, --buckets <buckets>       Name of bucket(s)with a comma separator if more than one
    -s, --start <start>           Start of time range
    -e --end <end>                End of time range
    -h, --host <host>             Host of the server
    -p, --port <port>             Port of the server
    --ssl                         Enable ssl
    -v, --verbose

A typical call to list metrics for a bucket demo to Utapi in a https enabled deployment would be

node bin/list_bucket_metrics -a myAccessKey -k mySecretKey -b demo -s 1476231300000 -h 127.0.0.1 -p 8100 --ssl

Both start and end times are time expressed as UNIX epoch timestamps expressed in milliseconds.

Keep in mind, since Utapi metrics are normalized to the nearest 15 min. interval, so start time and end time need to be in specific format as follows.

Start time

Start time needs to be normalized to the nearest 15 minute interval with seconds and milliseconds set to 0. So valid start timestamps would look something like 09:00:00:000, 09:15:00:000, 09:30:00:000 and 09:45:00:000.

For example

Date: Tue Oct 11 2016 17:35:25 GMT-0700 (PDT)

Unix timestamp (milliseconds): 1476232525320

Here's a typical JS method to get start timestamp

function getStartTimestamp(t) {
    const time = new Date(t);
    const minutes = time.getMinutes();
    const timestamp = time.setMinutes((minutes - minutes % 15), 0, 0);
    return timestamp;
},

This would format the start time timestamp to 1476231300000

End time

End time needs to be normalized to the nearest 15 minute end interval with seconds and milliseconds set to 59 and 999 respectively. So valid end timestamps would look something like 09:14:59:999, 09:29:59:999, 09:44:59:999 and 09:59:59:999.

Here's a typical JS method to get end timestamp

function getEndTimestamp(t) {
    const time = new Date(t);
    const minutes = time.getMinutes();
    const timestamp = time.setMinutes((minutes - minutes % 15) + 15, 0, -1);
    return timestamp;
}

This would format the end time timestamp to 1476233099999

Clone this wiki locally