usage

usage copied to clipboard

usage provides some basic tools to parse XML usage reports from Amazone Web Services. These reports contain detailed information about resources that you are using (and therefore will be billed for) for a given AWS account. There is no API available to download these usage reports from your AWS web console so you either have to download them manually or use a screen scraping approach as described here:

http://gist.github.com/421559

This software was written a while ago and has been languising on my hard drive ever since. I thought it might be useful to someone so I decided to put it up on github.

Using usage

Assuming you have an XML usage report ('report.xml'), you can parse that file like this:

In [1]: import usage.dataset

In [2]: ds = usage.dataset.DataSet()

In [3]: ds.load('report.xml')

In [4]: ds.keys() Out[4]: [u'AmazonEC2.ElasticIP-In.DataTransfer-Regional-Bytes', u'AmazonEC2.EBS:IO-Write.EBS:VolumeIOUsage', u'AmazonEC2.RunInstances.DataTransfer-Out-Bytes', u'AmazonEC2.EBS:IO-Read.EBS:VolumeIOUsage', u'AmazonEC2.PublicIP-In.DataTransfer-In-Bytes', u'AmazonEC2.RunInstances.BoxUsage', u'AmazonEC2.CreateVolume.EBS:VolumeUsage', u'AmazonEC2.InterZone-In.DataTransfer-Regional-Bytes', u'AmazonEC2.CreateSnapshot.EBS:SnapshotPutUsage', u'AmazonEC2.RunInstances.DataTransfer-In-Bytes', u'AmazonEC2.CreateSnapshot.EBS:SnapshotUsage', u'AmazonEC2.ElasticIP-Out.DataTransfer-Regional-Bytes', u'AmazonEC2.PublicIP-Out.DataTransfer-Regional-Bytes', u'AmazonEC2.AssociateAddress.ElasticIP:Remap', u'AmazonEC2.PublicIP-In.DataTransfer-Regional-Bytes', u'AmazonEC2.InterZone-Out.DataTransfer-Regional-Bytes']

As you can see, the DataSet class is a subclass of Python's dict object. Each value stored in the DataSet is of type usage.dimension.Dimension.

In [5]: dim = ds['AmazonEC2.RunInstances.BoxUsage']

In [6]: dim Out[6]: Dimension.AmazonEC2.RunInstances.BoxUsage

The Dimension object has a number of fields, such as the operation_name, the service, a start and end timestamp and a data attribute which is a list of actual data points pulled from the usage report. Each datum is of type usage.usage.Usage which has the attributes start, end and value:

In [9]: usage = dim.data[0]

In [10]: usage.start Out[10]: datetime.datetime(2009, 7, 9, 16, 0)

In [11]: usage.end Out[11]: datetime.datetime(2009, 7, 9, 17, 0)

In [12]: usage.value Out[12]: 4

There are some examples of reports that can be generated from the usage data in usage.reports. For example:

In [14]: import usage.reports

In [15]: ebs = usage.reports.EBSReports(ds)

In [16]: ebs.report() EBS Reports

59502 I/O Requests = 0.000000 545 SnapShot Puts = 0.000000 541129438500 SnapShot Storage = 81.150000

These reports are only examples and may not produce accurate data at this point but they might prove useful in developing your own reports.