stackdriver / collectd Goto Github PK

Stackdriver team,

The mongodb collectd plugin is currently not being shipped in EPEL and I believe that's related to the fact the "official" module hasn't seen an update since 2012 (according to [1]). According to [2] a fixed and confirmed working plugin (based on [1]) has been developed but the contributions were never submitted back upstream to be then consumed by the major Linux distributions out there. Would it be possible for this to happen in order for collectd to finally have an officially supported mongodb plugin?

Thanks!

[1] https://collectd.org/wiki/index.php/Plugin:MongoDB
[2] https://cloud.google.com/monitoring/agent/plugins/mongodb

Please let me know if this should be asked elsewhere.

I tried to install the monitoring agent based on instructions from the official docs but it turns out my OS is not supported; this is correctly reflected on the list of supported OSes, where Debian 9 "Stretch" is listed but no later versions are available.

Is there known incompatibility with Debian 10 "buster", or has an official release just not yet been created?

• Version of collectd: 6.1.1
• Operating system / distribution: Linux / Arch

Expected behavior

No error.

Actual behavior

The collectd program print the following error message,

Jan 14 14:43:05 <hostname> collectd[702017]: write_gcm: Server response (CollectdTimeseriesRequest) contains errors:
                                                 {
                                                   "summary": {
                                                     "totalPointCount": 399,
                                                     "successPointCount": 239,
                                                     "errors": [
                                                       {
                                                         "status": {
                                                           "code": 3,
                                                           "message": "Points must be written in order. One or more of the points specified had an older start time than the most recent point."
                                                         },
                                                         "pointCount": 160
                                                       }
                                                     ]
                                                   }
                                                 }

Steps to reproduce

• Run the collectd 6.1.1 in Linux VM
• Monitor the collectd logs using journalctl -f -u stackdriver-collectd
• Wait until the above error printed

• Version of collectd: 6.1.1
• Operating system / distribution: archlinux 5.10.16-arch1-1

Expected behavior

yay -S stackdriver-collectd should install

Actual behavior

Build/install fails with src/sensors.c:165:2: error: #error "This version of libsensors is not supported yet. Please report this " "as bug."

Sensors version:

> yay -Q lm_sensors          
lm_sensors 1:3.6.0.r41.g31d1f125-1

Steps to reproduce

• have lm_sensors 3.6.0 installed? (in another arch instance, collectd built fine. lm_sensors was not installed on the other instance)
• try to build collectd?

• Version of collectd: Whatever ships with stackdriver-agent-5.5.2-384.el7.x86_64 which I beleive is collectd 5.5.2.1446.g4be049e+
• Operating system / distribution: CentOS Linux release 7.6.1810 (Core)

Expected behavior

I expect the plugin configured as indicated here Google Raw StackDriver Varnish Config to actually load.

Actual behavior

systemctl restart stackdriver-agent.service
Job for stackdriver-agent.service failed because the control process exited with error code. See "systemctl status stackdriver-agent.service" and "journalctl -xe" for details.

journalctl -xe:
May 28 23:05:10 varnish-001 collectd[31561]: plugin_load: plugin "aggregation" successfully loaded.
May 28 23:05:10 varnish-001 collectd[31561]: lt_dlopen ("/opt/stackdriver/collectd/lib64/collectd/varnish.so") failed: file not found. The most common cause for this problem is missing dependencies. Use ldd(1) to check the dependencies of the plug
May 28 23:05:10 varnish-001 collectd[31561]: plugin_load: Load plugin "varnish" failed with status 1.
May 28 23:05:10 varnish-001 systemd[1]: stackdriver-agent.service: control process exited, code=exited status=1

ldd /opt/stackdriver/collectd/lib64/collectd/varnish.so
/opt/stackdriver/collectd/lib64/collectd/varnish.so: /lib64/libvarnishapi.so.1: version `LIBVARNISHAPI_1.0' not found (required by /opt/stackdriver/collectd/lib64/collectd/varnish.so)
linux-vdso.so.1 => (0x00007ffc10fdc000)
libvarnishapi.so.1 => /lib64/libvarnishapi.so.1 (0x00007fac4a0d3000)
libc.so.6 => /lib64/libc.so.6 (0x00007fac49d06000)
libpcre.so.1 => /lib64/libpcre.so.1 (0x00007fac49aa4000)
librt.so.1 => /lib64/librt.so.1 (0x00007fac4989c000)
libm.so.6 => /lib64/libm.so.6 (0x00007fac4959a000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fac4937e000)
/lib64/ld-linux-x86-64.so.2 (0x00007fac4a502000)

However libvarnishapi.so.1 is there as a sym link to the actual library, but the compiled varnish.so library from stackdriver does not believe it's the correct version.

Steps to reproduce

• Install Varnish 5.2.1
• Install Stack Driver 5.5.2
• Put the above referenced config file in collectd.d directory
• Restart stackdriver service
• Follow the steps laid out above to debug

• Version of collectd: GCM-46
• Operating system / distribution:
  Windows 2019

Expected behavior

The system user does not have permission to the disk created by the drive file stream. So the disk capacity is zero for the system user. [disk used / zero].
All disks that capacity is zero should not be considered.

Actual behavior

Error: Exception collecting Disk metrics: ZeroDivisionError('float division by zero',)

Steps to reproduce

• step 1 - Install and set up Google Drive File Stream
• step 2 - Disk utilization error

• Version of collectd: 6.1.3-1
• Operating system / distribution: CentOS Linux release 8.3.2011

Expected behavior

No errors with default configuration on virtual machine instances without swap.

Actual behavior

Constantly repeating errors in the journal (e.g. systemctl status stackdriver-agent or journalctl -xe) with the first line in red:

write_gcm: can not take infinite value
collectd[1520]: write_gcm: wg_typed_value_create_from_value_t_inline failed for swap/percent/value! Continuing.

Steps to reproduce

1. Install Cloud Monitoring agent following these instructions.
2. Check agent status (e.g. systemctl status stackdriver-agent) or for errors in general (e.g. journalctl -xe).

The workaround is of course to disable the swap plugin, but having virtual machine instances without swap is common nowadays, so there should not be error at all.

I'm experiencing an issue with the Stackdriver monitoring agent for VMs running in Google Cloud. Everything used to work as expected until yesterday (Dec, 1st).

Metrics are not sent to the Stackdriver Monitor console in Google Cloud. The error is reported below. Any ideas on what could be wrong? No changes were made.

• Version of collectd: 5.5.2-382.stretch
• Operating system / distribution: Debian Stretch 9.5

Expected behavior

Metrics are published to Stackdriver

Actual behavior

Metrics are not being sent to Stackdriver
Error in logs:

Dec 02 20:08:46 test collectd[29217]: write_gcm: Asking metadata server for auth token
Dec 02 20:08:46 test collectd[29217]: write_gcm: Error or buffer overflow when building auth_header
Dec 02 20:08:46 test collectd[29217]: write_gcm: wg_oauth2_get_auth_header failed.
Dec 02 20:08:46 test collectd[29217]: write_gcm: wg_transmit_unique_segment failed.
Dec 02 20:08:46 test collectd[29217]: write_gcm: wg_transmit_unique_segments failed. Flushing.

Steps to reproduce

Nothing specific, collectd has been up and running for months

In our k8s we are using stackdriver-agent in side containers to collect metrics from applications and send them as custom metrics to stackdriver through write_gcm.
It looks like in write_gcm the monitored resource is hardcoded to 'gce_instance', making not possible to have different metrics, one per pod, in case several replicas of a pod are running on the node.

• Version of collectd:5.5.2
• Operating system / distribution: CentOS Linux release 7.5/GKE

Expected behavior

Possible to set the monitored resource type.

Actual behavior

Monitored resource is hardcoded to 'gce_instance'

Steps to reproduce

Send any custom metric using stackdriver-agent and write_gcm plugin.

Hi ,
I have some issues with the latest stackdriver agent.

• Version : both 6.0.2-1.focal & 6.0.3-1.focal
• Operating system / distribution: Ubuntu 20.04 LTS

Expected behavior

Monitor the Disk utilization to alert when disks are nearly full.

Actual behavior

No data about Disk Utilization are pushed to the Monitoring system.
I have it on 2 Ubuntu 20.04 fresh install, but I tried to install 6.0.3-1.focal on an Ubuntu 18.04 and it's working fine.

I have this kind of logs every minute:

juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: can not take infinite value
juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: wg_typed_value_create_from_value_t_inline failed for swap/percent/value! Continuing.
juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: Unsuccessful HTTP request 400: {
                                                   "error": {
                                                     "code": 400,
                                                     "message": "Field timeSeries[0].points[0].interval.start_time had an invalid value of \"2020-06-17T22:56:44.709265-07:00\": Th>
                                                     "status": "INVALID_ARGUMENT"
                                                   }
                                                 }
juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: Error talking to the endpoint.
juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: wg_transmit_unique_segment failed.
juin 18 07:58:44 gce_instance collectd[13149]: write_gcm: wg_transmit_unique_segments failed. Flushing.

Steps to reproduce

• Create GCE Vm Ubuntu 20.04 LTS minimal
• Install and launch Stackdriver agent with official method: https://cloud.google.com/logging/docs/agent/installation
• On https://console.cloud.google.com/monitoring/alerting/policies/
  Add an alerting policy with:

• Resource type: GCE VM Instance
• Metrics: Disk Utilization
• Filters:
  • device: sda1 or your device if not sda1
  • state: used
  • instance_id: your instance id
• A numeric value un Threshold field

--> Message in the monitoring screen : No data is available for the selected time frame.