fluentd
fluentd copied to clipboard
metrics_agent plugin does not produce any metrics for non-output plugins
Describe the bug
I recently noticed that there is no usable data in output of the monitor_agent
plugin for any non-output plugins (except for retry_count)
To Reproduce docker-compose.yaml:
version: "3"
services:
fluentd:
image: fluent/fluentd:v1.11-2
volumes:
- ./fluent.conf:/fluentd/etc/fluent.conf
fluent.conf:
<source>
@type dummy
@id in_dummy
dummy {"hello":"world"}
auto_increment_key count
rate 1
tag dummy
</source>
<source>
@type monitor_agent
@id in_monitor
tag stats.fluentd
emit_interval 10
</source>
<match **>
@type stdout
@id out_stdout
</match>
output:
2020-10-13 20:41:07.002072200 +0000 dummy: {"hello":"world","count":19}
2020-10-13 20:41:07.500727200 +0000 stats.fluentd: {"plugin_id":"in_dummy","plugin_category":"input","type":"dummy","output_plugin":false,"retry_count":null}
2020-10-13 20:41:07.500727200 +0000 stats.fluentd: {"plugin_id":"in_monitor","plugin_category":"input","type":"monitor_agent","output_plugin":false,"retry_count":null}
2020-10-13 20:41:07.500727200 +0000 stats.fluentd: {"plugin_id":"out_stdout","plugin_category":"output","type":"stdout","output_plugin":true,"retry_count":0,"emit_records":25,"emit_count":23,"write_count":0,"rollback_count":0,"slow_flush_count":0,"flush_time_count":0}
Expected behavior
Expect to see number of records received/processed and other useful statistics on all emitted records, not just output one
Your Environment
For sake of testing, used docker image fluent/fluentd:v1.11-2
- but same seems to be happening in multiple versions
If you hit the problem with older fluentd version, try latest version first.
Your Configuration
(see above )
Your Error Log
(see above )
@mlasevich It seems that it is the intended behavior of fluentd. The response may change but it depends on plugin. ref. https://github.com/fluent/fluentd/blob/v1.11.2/lib/fluent/plugin/in_monitor_agent.rb#L238
FYI: you may able to get further information by http://127.0.0.1:24220/api/plugins.json?debug=1 query.
Yeah, seems that entire monitoring plugin is completely useless outside the output plugins. (and even that seems to be an afterthought, added later) - not really sure why it was created in the first place. Must have been a placeholder for something never completed :-(
This issue has been automatically marked as stale because it has been open 90 days with no activity. Remove stale label or comment or this issue will be closed in 30 days
Remove stale and add feature request. We need more feedback which metrics are needed for other plugin types.
Hello, for input plugins, is the retry_count always null or can it change during trouble? Is it worth monitoring?