gregdurham / collectd-kairosdb Goto Github PK

syslog reports error writing the collectd data to kairosdb.

Unhandled python exception in write callback: UnboundLocalError: local variable 'tags' referenced before assignment

Environment : debian 7.0
python : 2.7.3
collectd : 5.1.0.3

Having the option to remove the host from the metric name will make it easier to graph across hosts grouping by a host tag instead.

I'm happy to file a pull request with this change.

Hi,

I have a problem using your module, it looks like it doesn't work and if I try test the plugin with collectd -T I receive an segfault:

collectd[5153]: segfault at 98 ip 00007fb9d5cd5f8e sp 00007ffc69701e80 error 4 in libpython2.7.so.1.0[7fb9d5c3a000+178000]

Here is my Setup:

• OS: CentOS Linux release 7.1.1503 (Core)
• collectd Version: collectd-5.5.0-2.el7.x86_64
• installed Python packages:

python-libs-2.7.5-18.el7_1.1.i686
rpm-python-4.11.1-25.el7.x86_64
python-slip-dbus-0.4.0-2.el7.noarch
python-setuptools-0.9.8-3.el7.noarch
python-pycurl-7.19.0-17.el7.x86_64
python-2.7.5-18.el7_1.1.x86_64
python-urlgrabber-3.10-6.el7.noarch
python-decorator-3.4.0-3.el7.noarch
python-iniparse-0.4-9.el7.noarch
newt-python-0.52.15-4.el7.x86_64
python-configobj-4.7.2-7.el7.noarch
python-slip-0.4.0-2.el7.noarch
python-backports-ssl_match_hostname-3.4.0.2-4.el7.noarch
python-pyudev-0.15-6.el7.noarch
python-backports-1.0-8.el7.x86_64
python-libs-2.7.5-18.el7_1.1.x86_64
dbus-python-1.1.1-9.el7.x86_64
libselinux-python-2.2.2-6.el7.x86_64

• collectd.conf:

# cat /etc/collectd.conf |grep -v ^#|sed '/^$/d'
Hostname    "cent7"
BaseDir     "/var/lib/collectd"
PIDFile     "/var/run/collectd.pid"
PluginDir   "/usr/lib64/collectd"
TypesDB     "/usr/share/collectd/types.db"
AutoLoadPlugin true
CollectInternalStats true
Interval     10
ReadThreads     1
WriteThreads    1
LoadPlugin syslog
<Plugin syslog>
        LogLevel debug
</Plugin>
LoadPlugin cpu
LoadPlugin load
LoadPlugin memory
LoadPlugin python
<Plugin cpu>
  ReportByCpu true
  ReportByState true
  ValuesPercentage false
</Plugin>
Include "/etc/collectd.d"

• /etc/collectd.d/kairosdb.conf

cat /etc/collectd.d/kairosdb.conf |grep -v ^#|sed '/^$/d'
<LoadPlugin "python">
    Globals true
</LoadPlugin>
<Plugin "python">
    ModulePath "/usr/lib64/collectd/"
    Import "kairosdb_writer"
    <Module "kairosdb_writer">
        KairosDBHost "localhost"
        KairosDBPort 4242
        KairosDBProtocol "tcp"
        LowercaseMetricNames true
        TypesDB "/usr/share/collectd/types.db"
    </Module>
</Plugin>

Hope you could help.

Regards,
Sascha

I´m trying to send collect data to kairosdb, which seems to work - in a certain way.

I see all of my collectd metrics by checking my kairosdb URL
http://my.kairosdb.host:8080/api/v1/metricnames

{"results":["cpu.0.cpu.idle.value","interface.lo.if_errors.rx","processes.collectd.ps_count.processes","load.load.shortterm","load.load.longterm","processes.ps_state.stopped.value","interface.eth0.if_octets.tx","processes.fork_rate.value","vmem.vmpage_io.swap.out","aggregation.cpu-average.cpu.nice.value","kairosdb.jvm.total_memory","cpu.1.cpu.softirq.value","processes.collectd.ps_disk_ops.write","vmem.vmpage_number.active_file.value","aggregation.cpu-average.cpu.wait.value","memory.memory.buffered.value","memory.memory.cached.value","processes.collectd.ps_pagefaults.majflt","vmem.vmpage_faults.minflt","vmem.vmpage_number.anon_transparent_hugepages.value","vmem.vmpage_number.isolated_anon.value","interface.lo.if_errors.tx","processes.collectd.ps_disk_ops.read","kairosdb.datastore.query_collisions","aggregation.cpu-average.cpu.idle.value","vmem.vmpage_io.swap.in","interface.lo.if_octets.rx","users.users.value","processes.collectd.ps_data.value","processes.ps_state.zombies.value","processes.collectd.ps_code.value","vmem.vmpage_number.free_pages.value","uptime.uptime.value","vmem.vmpage_number.mapped.value","processes.collectd.ps_count.threads","kairosdb.jvm.thread_count","cpu.0.cpu.softirq.value","vmem.vmpage_number.inactive_file.value","aggregation.cpu-average.cpu.softirq.value","aggregation.cpu-average.cpu.system.value","vmem.vmpage_number.slab_reclaimable.value","processes.collectd.ps_disk_octets.write","vmem.vmpage_number.file_pages.value","vmem.vmpage_number.slab_unreclaimable.value","cpu.0.cpu.steal.value","cpu.1.cpu.steal.value","aggregation.cpu-sum.cpu.system.value","vmem.vmpage_number.active_anon.value","swap.swap.free.value","kairosdb.jvm.max_memory","processes.collectd.ps_disk_octets.read","processes.ps_state.running.value","kairosdb.jvm.free_memory","cpu.0.cpu.interrupt.value","processes.collectd.ps_cputime.user","memory.memory.used.value","interface.lo.if_octets.tx","aggregation.cpu-sum.cpu.softirq.value","aggregation.cpu-sum.cpu.user.value","interface.eth0.if_packets.rx","interface.eth0.if_errors.rx","kairosdb.protocol.http_request_count","processes.collectd.ps_stacksize.value","vmem.vmpage_io.memory.in","interface.lo.if_packets.rx","cpu.1.cpu.user.value","aggregation.cpu-sum.cpu.steal.value","aggregation.cpu-sum.cpu.nice.value","cpu.0.cpu.system.value","swap.swap_io.in.value","vmem.vmpage_number.mlock.value","vmem.vmpage_number.kernel_stack.value","swap.swap.cached.value","vmem.vmpage_number.inactive_anon.value","cpu.1.cpu.nice.value","vmem.vmpage_faults.majflt","aggregation.cpu-sum.cpu.interrupt.value","vmem.vmpage_number.anon_pages.value","processes.collectd.ps_pagefaults.minflt","vmem.vmpage_number.unstable.value","vmem.vmpage_number.vmscan_write.value","aggregation.cpu-sum.cpu.wait.value","processes.ps_state.sleeping.value","cpu.0.cpu.user.value","kairosdb.metric_counters","swap.swap_io.out.value","contextswitch.contextswitch.value","cpu.1.cpu.system.value","cpu.1.cpu.wait.value","vmem.vmpage_number.page_table_pages.value","interface.eth0.if_packets.tx","interface.eth0.if_errors.tx","cpu.0.cpu.nice.value","vmem.vmpage_number.writeback_temp.value","vmem.vmpage_number.writeback.value","interface.eth0.if_octets.rx","aggregation.cpu-sum.cpu.idle.value","interface.lo.if_packets.tx","load.load.midterm","processes.collectd.ps_vm.value","processes.ps_state.paging.value","vmem.vmpage_number.dirty.value","processes.collectd.ps_rss.value","cpu.1.cpu.idle.value","cpu.0.cpu.wait.value","vmem.vmpage_number.bounce.value","processes.ps_state.blocked.value","vmem.vmpage_number.shmem.value","vmem.vmpage_number.isolated_file.value","kairosdb.protocol.telnet_request_count","aggregation.cpu-average.cpu.steal.value","memory.memory.free.value","vmem.vmpage_number.unevictable.value","aggregation.cpu-average.cpu.user.value","processes.collectd.ps_cputime.syst","cpu.1.cpu.interrupt.value","aggregation.cpu-average.cpu.interrupt.value","vmem.vmpage_io.memory.out","swap.swap.used.value"]}

but I cannot see my hostname, for example "safara".
I was thinking that all the metrics are prefixed with my hostname?

thanks
,
Stefan

<LoadPlugin "python">
Globals true

<Plugin "python">
ModulePath "/usr/lib64/collectd/"

Import "kairosdb_writer"

<Module "kairosdb_writer">
    KairosDBHost "my.kairosdb.host"
    KairosDBPort 4242
    KairosDBProtocol "tcp"
    LowercaseMetricNames true
    TypesDB "/usr/share/collectd/types.db"
    Tags "host=safara.my.fqdn" "role=safara"
</Module>