vimeo/nagios-cloudwatch-plugin

AWS CloudWatch check Nagios plugin

nagios-cloudwatch-plugin's Issues

nagios-cloudwatch-plugin 0.2.6 throws "CLOUDWATCHMETRIC UNKNOWN: AttributeError"

I have tried installing the Nagios to Cloudwatch API given on below link:

https://pypi.python.org/pypi/nagios-cloudwatch-plugin/0.2.6

However, while executing the below script to check the output, i get below error:

./check_cloudwatch.py -R us-east-2 -n AWS/EC2 -m CPUUtilization -p 600 -d InstanceId=i-0305354ba3f9cd1b1 -w 0:75 -c 0:90
CLOUDWATCHMETRIC UNKNOWN: AttributeError: 'NoneType' object has no attribute 'get_metric_statistics'

I'm using following version of phython boto:
python-boto-2.34.0-6.el6.noarch

The credentials for the AWS cloudwatch access are stored in /etc/boto.cfg file

Can someone please check why this error is thrown ? Appreciate a quick response.

Perf counters do not follow standard

https://nagios-plugins.org/doc/guidelines.html

Can add something like this:

    perfstat_compatible_uom = {'Percent':'%','Bytes':'B','Count':'','Megabytes':'MB'}

    if stat['Unit'] in perfstat_compatible_uom:
        stat['Unit'] = perfstat_compatible_uom[stat['Unit']]

    return [nagiosplugin.Metric('cloudwatchmetric', stat[self.statistic], stat['Unit'])]

RDS call not working.

Does it produce RDS cloudwatch?

is there a sample call if you have tried it?

Empty but valid AWS result causes a failure

We use this plugin to check our AWS SQS dead queues (0 messages = good). It seems like Amazon responds with an empty result instead of a result with 0-values. In this case, the plugin gives us the following error:

WARNING:nagiosplugin.check:resource CloudWatchMetric did not produce any metric
CLOUDWATCHMETRIC UNKNOWN - no check results
resource CloudWatchMetric did not produce any metric (check.py:78)

The problem is, that this triggers a PROBLEM in our Icinga installation.

Is there a trick to avoid this error message? Do we need to set or unset a configuration flag? Do you know if there is a way to tell Amazon to give use 0-values in case of no data?

Maintenance status of this project

Is this project still maintained? Would you be willing to accept PRs for tests and argparser bugfixes?

Add possibility to use profiles described in boto config file

I have two separate accounts for live and staging. I described both in boto config file and added option to use profile name if we need it. Default value for profile name is 'default'.

Example

I'm struggling to get this to work. An example of the syntax for someone unfamiliar with CloudWatch would be really helpful. e.g. given an EC2 instance ID, what would the command line look like to get CPUUtilization for that instance?