Telegraf agent not polling the input as per the configured interval

[shansudhakar]

Directions

GitHub Issues are reserved for actionable bug reports and feature requests.
General questions should be asked at the InfluxData Community site.

Before opening an issue, search for similar bug reports or feature requests on GitHub Issues.
If no similar issue can be found, fill out either the "Bug Report" or the "Feature Request" section below.
Erase the other section and everything on and above this line.

Please note, the quickest way to fix a bug is to open a Pull Request.

Bug report

Relevant telegraf.conf:

[agent]
interval = "60s"
round_interval = true
metric_batch_size = 10000
metric_buffer_limit = 100000
collection_jitter = "0s"
flush_jitter = "0s"
precision = ""

[[outputs.file]]
files = ["stdout", "/tmp/metrics.out"]
data_format = "influx"

[[outputs.kafka]]
brokers = ["measure-kafka-tgt-ttc.target.com:9092"]
required_acks = 0
topic = "metric-influx-tgt-ttc-prod-platform-engineering-mft"

# Stream and parse log file(s).

[[inputs.logparser]]
files = ["/apps/mft/shared/logs/MFT_CheckNode.log"]
from_beginning = true

[inputs.logparser.grok]
patterns = ["%{LastHeartBeat}"]
measurement = "mft_lastbeat_prod"
custom_patterns = '''
LastHeartBeat %{TIMESTAMP_ISO8601:timestamp}|%{WORD:node:tag}.+|%{BASE10NUM:cstatus1:int}.* %{BASE10NUM:cstatus2:int}.* %{BASE10NUM:cstatus3:int}
'''

System info:

telegraf]# telegraf -version
Telegraf v1.4.3 (git: release-1.4 b473b6a)

telegraf]# cat /etc/redhat-release
Red Hat Enterprise Linux Server release 7.1 (Maipo)

Steps to reproduce:

1. Provide input and write the output to log/kafka topic
2. Restart the telegraf agent

Expected behavior:

Poll every 60s of the input log provided

Actual behavior:

Its reading the input only when telegraf agent restarted, that too with "from_beginning = true"

Additional info:

"from_beginning = false" not at all working to get the delta from the input log file

[Include gist of relevant config, logs, etc.]
[<script src="https://gist.github.com/shansudhakar/c74bf66797d86cc83f4103033af9ada3.js"></script>]
[<script src="https://gist.github.com/shansudhakar/e3eacb043d5c50fe656effbe54b0ac36.js"></script>]
[<script src="https://gist.github.com/shansudhakar/9b93fc90ab0aba2952c97ac7251e4bdb.js"></script>]

Feature Request

Opening a feature request kicks off a discussion.

Proposal:

Current behavior:

Desired behavior:

Use case: [Why is this important (helps with prioritizing requests)]

Read multiple logs at a frequency and display status of infra on grafana dashboard

[danielnelson]

The logparser input is a so called service input, which are event driven. Instead of using the interval these inputs generate metrics when events occur, in the case of logparser this is when the logfile is appended too.

If you have from_beginning = true, on startup the entire file will be read and processed, otherwise only newly added lines will be detected.

[shansudhakar]

@danielnelson I understand the functionality of the logparser, however the issue is agent/plugin not reading the logfile whenever its appended with new entries. It reads only with "from_beginning = true" and agent is restarted. Our requirement is to have "from_beginning = false" and read/process the appended entries into the log file input.

[danielnelson]

If logparser is not picking up new lines, but they are parsed when starting up with from_beginning = true then it is a bug. We do know of one bug with change detection #2847, maybe this is what you are seeing?

[shansudhakar]

@danielnelson #2847 seems to be a different issue with new file not detected after refresh/rotate.

[danielnelson]

Are you able to reliably reproduce this issue? If so can you try to create a list of the smallest number of steps to reproduce?

BTW, when 1.5 is released or with the nightlies you can also use method = "poll" to poll the file for changes instead of using inotify.

[shansudhakar]

which is the correct config file to use:
/etc/telegraf/telegraf.d/default_outputs.conf or /etc/telegraf/telegraf.conf?

[danielnelson]

/etc/telegraf/telegraf.conf is the default config, the other one must be something custom to your site. Don't forget that all files in /etc/telegraf/telegraf.d are loaded by default.

[danielnelson]

Closing as I don't think there is any action to take.