Check_AWS_CloudWatch_metrics

Dear Coder,
First of all, alot of thanks for putting this great plugin. It has removed the depedency for installing nrpe on each instance. I am able to install all depedencies. Now when i tried to run the script i am having this error. Can you please tell me what i am missing here.

Error occured while retrieving and decrypting credentials for instance i-XXXXXXXXX on Amazon Server: ec2-XXXXXXXXX-compute-1.amazonaws.com: key must be 16 bytes

Change: cipher = OpenSSL::Cipher::Cipher.new('bf-cbc') TO: cipher = OpenSSL::Cipher::AES256.new(:CBC)
In check_cloudwatch_status-6.4.44.rb and encrypt_credentials.rb

If you are going to use those scripts and your Ruby version is 2.4x, make sure do some modifications on the script:

1. Change "cipher = OpenSSL::Cipher::Cipher.new('bf-cbc')" into ''cipher = OpenSSL::Cipher.new('bf-cbc')"

2. Change "cipher = OpenSSL::Cipher::Cipher.new('bf-cbc')" into "cipher = OpenSSL::Cipher::Cipher.new('aes-256-bc')".

Are we able to access CloudWatch Metrics in Nagios via a proxy ?
If, not what changes do we need to make ?

With all up to date, every command comes with a

[fog][DEPRECATION] describe_load_balancers with String is deprecated, use all('LoadBalancerNames' => []) instead (./check_cloudwatch_status.rb:386:in `')

(which can be safely ignored with a simple 2>/dev/null).

Problem is, ELB's RequestCount does not work, it always returns 1:

# ./check_cloudwatch_status.rb -f ec2_credentials.cfg -i XXXX-elb -a XXXX-elb-XXXX.eu-west-1.elb.amazonaws.com -L RequestCount 2>/dev/null
CloudWatch Metric: RequestCount, Average: 1.00, Maximum: 1.00, Minimum: 1.00|metric_average=1.00 metric_maximum=1.00 metric_minimum=1.00

The real RequestCount is by the thousands.

This seems like a really good addition to our plugins, but I can't seem to make it work within Nagios. Reports metrics as expected from the command line, from any directory and as the nagios user, but I can only get a "null" response within nagios. I can't figure out what I'm missing.
I even tried the "--help" option as an argument and still only get a "null" return status.
Anybody else had a similar experience?

thanks

I've got this running, however this is without the -w and -c switches as the help and documentation are unclear whether they are percentages or absolute values. I'm wondering if this can be cleared up as I'm trying to set a threshold on RDS FreeStorageSpace

/tmp/test/check_cloudwatch_status-2.5.07.rb:382:in `+': no implicit conversion of Fog::AWS::RDS::Error into String (TypeError)
from /tmp/test/check_cloudwatch_status-2.5.07.rb:382:in `rescue in '
from /tmp/test/check_cloudwatch_status-2.5.07.rb:336:in `'

I am getting above error. Please help me to resolve it.

This would be so useful but using the public address rather than just specify region means that instances without public addresses cannot be monitored without hacking the code. This is an incredibly basic and obvious feature and has been mentioned some time ago so it's disappointing it still isn't implemented

Hello,
I am using Opsview4.5 built on Nagios Core 4.0.20130912. My ruby version is 2.2.

I have added some AWS hosts using default check_cloudwatch_status (first applied these: http://docs.opsview.com/doku.php?id=opsview4.5:opspacks:templates:amazoncloudwatch) script provided in Opsview and got following error:

/usr/local/nagios/libexec/check_cloudwatch_status:330:in `+': no implicit conversion of NameError into String (TypeError)
from /usr/local/nagios/libexec/check_cloudwatch_status:330:in `rescue in '
from /usr/local/nagios/libexec/check_cloudwatch_status:321:in `'


Then I decided to upgrade to the latest version. The error message changed but Opsview still cannot run the check. It quits with following error:

/usr/local/rvm/rubies/ruby-2.2.0/lib/ruby/site_ruby/2.2.0/rubygems/core_ext/kernel_require.rb:54:in `require': cannot load such file -- fog (LoadError)
from /usr/local/rvm/rubies/ruby-2.2.0/lib/ruby/site_ruby/2.2.0/rubygems/core_ext/kernel_require.rb:54:in `require'
from /usr/local/nagios/libexec/check_cloudwatch_status:12:in `block in '
from /usr/local/nagios/libexec/check_cloudwatch_status:12:in `each'
from /usr/local/nagios/libexec/check_cloudwatch_status:12:in `'


I have previously installed fog using "gem install fog" command, however current ruby executable cannot "see" fog package (sorry, I am not familiar with ruby).

Here is a sample check command which returns errors above:
sudo -u nagios /usr/local/nagios/libexec/check_cloudwatch_status -a dpesmas01q.n11.local -i "i-XXXXaab1" -C CPUUtilization -f /usr/local/nagios/etc/aws_credentials.cfg


Can you please help?

First of all, the following error:

`+': no implicit conversion of NoMethodError into String (TypeError)

can generally be fixed by replacing:

abcd" + e

with:

abcd #{e}"

The issue i have with this plugin is that it's not very clear how to set --critical and --warning threshold for metrics which are not percentages, but are absolute values.

Let's say, i want to trigger WARN if FreeStorageSpace reaches 10%. How do i do that?

Looking at the code, it seems that threshold can only be an absolute value. But what if I don't know what total available disk size for RDS really is? Even more, what if i want to use the same Nagios service for multiple RDS hosts where disk size might be different (and i still only want to trigger WARN using some percentage threshold)?

Is it possible to use percentage threshold metrics others than CPUUtilization?

I am getting similar error:
/usr/local/nagios/libexec/check_cloudwatch_status.rb:454:in `+': no implicit conversion of AWS::Error into String (TypeError)
from /usr/local/nagios/libexec/check_cloudwatch_status.rb:454:in `rescue in '
from /usr/local/nagios/libexec/check_cloudwatch_status.rb:408:in `'
Please help!

When I ran the script check_cloudwatch_status-2.1.77.rb, I get the following error:

check_cloudwatch_status-2.1.77.rb:524:in `+': no implicit conversion of NoMethodError into String (TypeError)
from check_cloudwatch_status-2.1.77.rb:524:in `rescue in '
from check_cloudwatch_status-2.1.77.rb:512:in `


Please help!

For ELB and RDS you don't need to define the instance_id so updated code to take hostname and shorten it for the elb/rds name. You can then use $HOSTNAME$ in the Nagios service check

check_cloudwatch_status.rb! YOURPATH/nagios/plugins/check_cloudwatch_status.rb -a $HOSTNAME$ -f ec2_credentials.cfg -L RequestCount -S Sum

line 348 :-

if namespace.eql?(AWS_NAMESPACE_EC2)
dimensions = "InstanceId=#{instance_id}"
elsif namespace.eql?(AWS_NAMESPACE_RDS)
# Strip off the name from the address
instance_id, *restofvalues = address.split('.')
instance_id = instance_id.split('-')[0...-1].join('-')
dimensions = "DBInstanceIdentifier=#{instance_id}"
elsif namespace.eql?(AWS_NAMESPACE_ELB)
# Strip off the name Load-Balancer-Name from the address Load-Balancer-Name-123456789.elb.amazon.com
instance_id, *restofvalues = address.split('.')
instance_id = instance_id.split('-')[0...-1].join('-')
dimensions = "LoadBalancerName=#{instance_id}"
end

Hi all,

I tested this script quite a bit recently, and offer the following comments:

* Very good integration with AWS once you get the grip on how to configure it.

* Instead of using the "Amazon Instance Address" with the -a/--address parameter, just use the region (i.e us-west-1 ) instead, since the script only use the address to parse out the region.

* I have offered the author a patch to enable http proxy usage, and to optionally allow credentials to be saved into a config.yml instead of the encrypted credentials file.

I have it working however it's funky depending on Cloudwatch's metrics.

It only works for when the measured metric from Cloudwatch is numerically greater than the warning or critical threshold metrics. So it works great for CPU since if it measures 80 from RDS Cloudwatch that meants it's at 80% use. However when measuring RDS FreeStorageSpace it won't work the same. FreeStorageSpace measures the empty space in RDS. This script won't work when you want to alert warning at 20GB free and then critical at 10GB free it will never trigger.

What would be ideal is if CloudWatch could measure UsedStorageSpace as well as FreeStorageSpace. Is there any way this script can measure the inverse?

If I could alert warning when RDS is 80% USED and 90% USED I think this script would be even more powerful.

I have incorporated this plugin with our existing nagios setup and it has been very useful.
It works exactly as expected and makes it easier to have all the metrics monitored.
Very easy to setup. Needed a little bit of testing etc.
Here are the example syntaxes for all the AWS components.

For ELB
RUBYOPT=rubygems ruby ~/check_cloudwatch_status.rb -a checkcheck-892345671.us-east-1.elb.amazonaws.com -i checkcheck -f ec2_credentials.cfg -L RequestCount -S Sum

As I was informed by Fred, you can also use -S Sum,Maximum

For RDS

RUBYOPT=rubygems ruby ~/check_cloudwatch_status.rb -a checkcheck-892345671.us-east-1.rds.amazonaws.com -i checkcheck -f ec2_credentials.cfg -D DiskQueueDepth

EC2 example is as given in the description.

Thank you Fred.

Hi.

First things first, Thanks for Check_AWS_CloudWatch_metrics Nagios Pluggin. The impementation and configuration was really quick and simple. It eased a lot of my work

This is the output I get

CloudWatch Metric: CPUUtilization, Average: 64.81, Maximum: 75.50, Minimum: 40.51 | metric_average=64.81 metric_maximum=75.50 metric_minimum=40.51


The only problem i had was to understanding the correctness of the output. When the pluggin says
CPUUtilization, Average: 64.81, Maximum: 75.50, Minimum: 40.51
Is it the average CPUUtilization, considering the CPU utilization of the past hour or the past day. Same in the case of Minimum and Maximum CPUUtilization. So if you could help me understand the output. it would be of great help to me

THANK YOU
Arun

great plugin, but i have a trouble. it works perfectly in terminal but when i configure it in nagios always show null in status information. what is going wrong?. thx

gem install CloudyScripts
mkdir /etc/cloutomate/
ssh-keygen -f /etc/cloutomate/cloutomate.pem
Enter blank passphrase
chown nagios:nagios cloutomate.pem

I was able to get this implemented VERY quickly.

My only question is this: How would you set this up if you are using two different regions? I tried just making a second copy of the file, but this appears to not work from within Nagios, though I can run the ruby script from the command line and get proper output.

Hey the plugin is great,
i'm using it for my vpc instances - for the address argument i'm using the aws region and it works perfectly.

Please fix the hard-coded .pem file to be configurable ...
One can create the rsa files via openssl genrsa ...

Thanks a lot

A very useful plug-in that allows us to add CloudWatch metrics to our existing monitoring dashboard.

This is excellent script to monitor EC2 and RDS .The
best feature is we can check free storage space in RDS .

Need to install amazon-ec2 gem, but other than that, setup is easy and it works great! Very good plugin for nagios to monitor AWS metrics

I needed this more for ELB checking then ec2 checking, does the ec2 checking quite well from what I've seen and relatively easy to set up.

However I will ask, is there a way of finegraining elb metrics? At the moment it just uses the entire region, can the -I flag be used for the elb dns entry?

Cheers.

Could not use the encrypt part
There is a hard code to /etc/cloutomate/cloutomate.pem what is that?
I always get the error:
Exception `NameError' at encrypt_credentials.rb:91 - uninitialized constant Digest::SHA256

Can you help me out?

Great plugin. I now can check my amazon instance with nagios ! The ability to see what's going on on all my instances is just great.

just to let people know it needs the amazon-ec2 gem installed to get it working as advertised and a little bit of work done to the rds monitoring to account for api changes