Host tags do not seem to be applied to the metrics as expected. They are clearly part of the host metadata being sent when the agent comes up:
https://github.com/DataDog/datadog-unix-agent/blob/master/metadata/metadata.py#L61
https://github.com/DataDog/datadog-unix-agent/blob/master/agent.py#L45

Expected:
Host tags to be associated with every tag coming from the host.

Actual:
Only metric tags available for aggregation on metrics.

OS - AIX 7200-05-03-2136
====================prtconf=================
System Model: IBM,9117-MMC
Processor Type: PowerPC_POWER7
Processor Implementation Mode: POWER 7
Processor Version: PV_7_Compat
Number Of Processors: 45
Processor Clock Speed: 3304 MHz
CPU Type: 64-bit
Kernel Type: 64-bit
LPAR Info: 12 aibdbsrv06
Memory Size: 256000 MB
Good Memory Size: 256000 MB
Platform Firmware level: AM770_126
Firmware Version: IBM,AM770_126
Console Login: enable
Auto Restart: true
Full Core: false
NX Crypto Acceleration: Not Capable
In-Core Crypto Acceleration: Not Capable

Network Information
Host Name: aibdbsrv06
IP Address: 10.10.10.26
Sub Netmask: 255.255.255.0
Gateway: 10.10.10.1
Name Server: 10.10.12.2
Domain Name: awash.local

Paging Space Information
Total Paging Space: 4096MB
Percent Used: 10%

Volume Groups Information

Active VGs

fedbcg:
PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION
hdisk1 active 999 0 00..00..00..00..00
hdisk2 active 999 0 00..00..00..00..00
hdisk4 active 999 0 00..00..00..00..00
hdisk5 active 999 0 00..00..00..00..00
hdisk6 active 999 0 00..00..00..00..00
hdisk7 active 999 0 00..00..00..00..00
hdisk8 active 999 0 00..00..00..00..00
hdisk9 active 999 0 00..00..00..00..00
hdisk10 active 999 0 00..00..00..00..00
hdisk11 active 999 0 00..00..00..00..00
hdisk12 active 999 0 00..00..00..00..00
hdisk13 active 999 0 00..00..00..00..00
hdisk14 active 999 0 00..00..00..00..00
hdisk15 active 999 22 00..00..00..00..22

db2logs:
PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION
hdisk0 active 999 0 00..00..00..00..00
hdisk16 active 599 1 00..00..00..00..01

rootvg:
PV_NAME PV STATE TOTAL PPs FREE PPs FREE DISTRIBUTION
hdisk3 active 831 342 112..58..00..165..07

INSTALLED RESOURCE LIST

The following resources are installed on the machine.
+/- = Added or deleted from Resource List.

• = Diagnostic support not available.

Model Architecture: chrp
Model Implementation: Multiple Processor, PCI bus

• sys0 System Object
• sysplanar0 System Planar

• vio0 Virtual I/O Bus
• vscsi0 U9117.MMC.064AC67-V12-C20-T1 Virtual SCSI Client Adapter
• cd1 U9117.MMC.064AC67-V12-C20-T1-L8200000000000000 Virtual SCSI Optical Served by VIO Server
• cd0 U9117.MMC.064AC67-V12-C20-T1-L8100000000000000 Virtual SCSI Optical Served by VIO Server
• ent1 U9117.MMC.064AC67-V12-C3-T1 Virtual I/O Ethernet Adapter (l-lan)
• ent0 U9117.MMC.064AC67-V12-C2-T1 Virtual I/O Ethernet Adapter (l-lan)
• vsa0 U9117.MMC.064AC67-V12-C0 LPAR Virtual Serial Adapter
• vty0 U9117.MMC.064AC67-V12-C0-L0 Asynchronous Terminal

• fcs0 U9117.MMC.064AC67-V12-C30-T1 Virtual Fibre Channel Client Adapter

• fscsi0 U9117.MMC.064AC67-V12-C30-T1 FC SCSI I/O Controller Protocol Device

Hi,
I am not sure, if this is the correct repository to report this issue.

When I execute the script (I got it from https://s3.amazonaws.com/dd-agent/scripts/install_script.sh) with the environment variable DD_HOST_TAGS=foo:bar it seems that the resulting configuration file datadog.yaml is broken.

I traced the problem to install_script.sh#L239-L243 where the command

"sed \"s|# tags:.*|tags: "$formatted_host_tags"/\" $CONF"

"uncomments" the correct part in datadog.yaml but unfortunately also some other part related to snmp configuration.

Since that other part is in the "middle of nowhere", where all surrounding config is still commented, this results in a configuration file that the datadog agent cannot parse.

It would be great if this issue could be fixed.

Cheers.

Perhaps hostnames that include underscores can be automatically replaced with hyphens in order for the agent to start?

The current workaround is to add a hostname parameter in the datadog.yaml configuration file that specified a hostname without underscores.

Here is the error message:

 > ./agent.py start
2018-08-09 12:24:20,918 | INFO | dd.agent | utils.pidfile(pidfile.py:36) | Pid file is: /tmp/datadog-unix-agent.pid
2018-08-09 12:24:20,919 | INFO | dd.agent | root(agent.py:227) | Start daemon
2018-08-09 12:24:20,919 | INFO | dd.agent | utils.daemon(daemon.py:95) | Starting...
2018-08-09 12:24:20,920 | DEBUG | dd.agent | utils.process(process.py:126) | Popen(['/bin/hostname', '-s'], stdout = -1) called
2018-08-09 12:24:20,931 | WARNING | dd.agent | utils.hostname(hostname.py:33) | Hostname: l490vp026_pub is not complying with RFC 1123
2018-08-09 12:24:20,931 | WARNING | dd.agent | utils.hostname(hostname.py:33) | Hostname: l490vp026_pub is not complying with RFC 1123
2018-08-09 12:24:20,931 | CRITICAL | dd.agent | utils.hostname(hostname.py:70) | Unable to reliably determine host name. You can define one in datadog.conf or in your hosts file
Traceback (most recent call last):
  File "./agent.py", line 248, in <module>
    sys.exit(main())
  File "./agent.py", line 228, in main
    agent.start(foreground=foreground)
  File "/opt/datadog/datadog-unix-agent/utils/daemon.py", line 114, in start
    self.run()
  File "./agent.py", line 133, in run
    hostname = get_hostname()
  File "/opt/datadog/datadog-unix-agent/utils/hostname.py", line 71, in get_hostname
    raise Exception('Unable to reliably determine host name. You can define one in datadog.conf or in your hosts file')
Exception: Unable to reliably determine host name. You can define one in datadog.conf or in your hosts file

Any Plans to migrate agents to Golang?

There are outstanding issues on PR #18 that should be addressed to possibly clean up the logic pertaining to the reporting, book-keeping of metric statistics.

The bulk of the PR is fine, however, so to avoid a continuous rebase I am going to proceed to merge it. The stats reporting is also non-critical and since it doesn't appear to be buggy (just sub-optimal, and maybe not the most maintainable).

This issue should be a reminder to revisit the relevant code and fixed if possible.

agent.log
agent_1.log