librato / collectd-librato Goto Github PK

we're adding this to our chef recipes, need this version to be easily accessible, thanks!

I am still pretty new to collectd. Playing around with this plugin I got an error:

Oct 29 12:26:18 titan collectd[20703]: Unhandled python exception in write callback: KeyError: 'floor_time_secs', file: collectd-librato.py,

I popped it in a try block:
# Floor measure time?
m_time = int(v.time)
try:
if config['floor_time_secs']:
m_time /= config['floor_time_secs']
m_time *= config['floor_time_secs']
except:
pass

This is around line 311.

I am also pretty new to open source collaboration. I can certainly edit the file and check in the change. I haven't done it before to someone else's repository, but, if this is a valid change I'd be happy to try it.

Best,
-greg

This and IncludeRegex would be mutually exclusive.

Collectd's configuration supports multiple TypesDB configuration directives, however it seems that the plugin is only aware of one.

Please see http://collectd.org/documentation/manpages/types.db.5.shtml#custom_types for further documentation of this.

Thank you! 🐮

Been attempting to deploy the collectd-librato plugin using the Chef cookbook as well. I am able to successfully run the cookbook and launch collectd, however it looks like urllib2 is reaching an IOError:

Mar 25 17:55:31 localhost.localdomain collectd[11012]: Collectd-Librato.py: Error when sending metrics Librato ([Errno 13] Permission denied)
Mar 25 17:56:01 localhost.localdomain collectd[11012]: Collectd-Librato.py: Error when sending metrics Librato ([Errno 13] Permission denied)

Here is my configuration for the plugin:

# This file autogenerated by Chef
# Do not edit, changes will be overwritten
<LoadPlugin python>
  Globals true
</LoadPlugin>

<Plugin python>
  ModulePath "/usr/lib64/collectd"
  ModulePath "/opt/collectd-librato-0.0.10/lib"
  Import "collectd-librato"

  <Module "collectd-librato">
    APIToken "#####"
    Email "######"
  </Module>
</Plugin>

I'm assuming this is a permissions issue, however unsure the files in question that would be causing this.

Collectd-librato uses the configured Hostname from the main collectd configuration as the metric source name. Add an option to collectd-librato's configuration section that allows the source name to be overridden.

Hello, here is a snippet of my collectd configuration:

    <LoadPlugin "python">
        Globals true
    </LoadPlugin>
    <Plugin "python">
        # collectd-librato.py is at /opt/collectd-librato-0.0.10/lib/collectd-librato.py
        ModulePath "/opt/collectd-librato-0.0.10/lib"
        Import "collectd-librato"
        <Module "collectd-librato">
            Email "<email>"
            APIToken "<token>"
            LowercaseMetricNames true
            MetricPrefix ""
            IncludeRegex "^cpu\..+$,^df\.mnt.+$,^df\.root\..+$,^memory\.memory\..+$,^ping\.ping_droprate\..+$,^swap\.swap\..+$"
        </Module>
    </Plugin>

For the most part this configuration is working as expected, however the swap.swap_io.in and swap.swap_io.out metrics are being not being filtered out.

I do not believe this to be a problem with the regex, as the following Python snippets properly filter out those metrics:

>>> import re
>>> "^cpu\..+$,^df\.mnt.+$,^df\.root\..+$,^memory\.memory\..+$,^ping\.ping_droprate\..+$,^swap\.swap\..+$".split(',')[5]
'^swap\\.swap\\..+$'
>>> re.match("^cpu\..+$,^df\.mnt.+$,^df\.root\..+$,^memory\.memory\..+$,^ping\.ping_droprate\..+$,^swap\.swap\..+$".split(',')[5], "swap.swap.cached")
<_sre.SRE_Match object at 0x10465d168>
>>> re.match("^cpu\..+$,^df\.mnt.+$,^df\.root\..+$,^memory\.memory\..+$,^ping\.ping_droprate\..+$,^swap\.swap\..+$".split(',')[5], "swap.swap_io.in")
>>> 

Any help would be greatly appreciated.

Hey Guys,

I got the plugin working but after a few hours it stops sending metrics and I have to restart collectd service.

When I restart collectd, I get this error below, but if I open the conf file and do anything like remove a line return or even some whitespace and save the file, collectd is able to restart without a problem and metrics start to enter librato. Any Ideas?

Restarting statistics collection and monitoring daemon: collectdParse error in file /etc/collectd/plugins/librato.conf', line 54 near': syntax error, unexpected $end, expecting EOL
yyparse returned error #1
configfile: Cannot read file /etc/collectd/plugins/librato.conf'. Parse error in file/etc/collectd/plugins/librato.conf', line 54 near ': syntax error, unexpected $end, expecting EOL yyparse returned error #1 configfile: Cannot read file/etc/collectd/plugins/librato.conf'.

librato.conf

<LoadPlugin "python">
Globals true

<Plugin "python">
    # collectd-librato.py is at /opt/collectd-librato-0.0.8/lib/collectd-librato.py
    ModulePath "/opt/collectd-librato-0.0.8/lib"

    Import "collectd-librato"

    <Module "collectd-librato">
        APIToken "<API TOKEN HERE>"
        Email    "[email protected]"
        Source "app01"
    </Module>
</Plugin>

running 0.0.2 of collectd-librato on debian 6, and python 2.6.6 getting the following error when checking my collectd config with the -t flag:

unhandled python exception in loading module: AttributeError: 'tuple' object has no attribute 'major'

I believe I've tracked down this issue to the following line:
https://github.com/librato/collectd-librato/blob/master/lib/collectd-librato.py#L117

if I run python REPL, and type sys.version_info, the tuple I get in py 2.6 looks like this:
(2, 6, 6, 'final', 0)

calling major on that results in an identical error message. I could patch this, but I am uncertain how you'd like to handle this scenario.