Process exporter metrics

A plugin based on Process Exporter to collect process level of metrics of system metrics

Prometheus Node exporter is a popular way to collect system level metrics from operating systems such as CPU, disk, network, and process statistics.

Fluent Bit 2.2 and later includes a process exporter plugin that builds off the Prometheus design to collect process level metrics without having to manage two separate processes or agents.

The Process Exporter Metrics plugin implements collecting of the various metrics available from the third party implementation of Prometheus Process Exporter and these will be expanded over time as needed.

All metrics including those collected with this plugin flow through a separate pipeline from logs and current filters don't operate on top of metrics. This plugin is only supported on Linux based operating systems as it uses the proc filesystem to access the relevant metrics. MacOS doesn't have the proc filesystem so this plugin won't work for it.

Configuration

Key	Description	Default
metrics	Specify which process level of metrics are collected from the host operating system. Actual values of metrics will be read from /proc when needed. context_switches, cpu, fd, io, memory, start_time, state, thread, and thread_wchan metrics depend on procfs.	cpu,io,memory,state,context_switches,fd,start_time,thread_wchan,thread
path.procfs	The mount point used to collect process information and metrics. Read-only permissions are enough.	/proc
process_exclude_pattern	Regular expression to determine which names of processes are excluded in the metrics produced by this plugin. It's not applied unless explicitly set.	NULL
process_include_pattern	Regular expression to determine which names of processes are included in the metrics produced by this plugin. It's applied for all process unless explicitly set.	.+
scrape_interval	The rate, in seconds, at which metrics are collected.	5

Available metrics

Name	Description
context_switches	Exposes context_switches statistics from /proc.
cpu	Exposes CPU statistics from /proc.
fd	Exposes file descriptors statistics from /proc.
io	Exposes I/O statistics from /proc.
memory	Exposes memory statistics from /proc.
start_time	Exposes start_time statistics from /proc.
state	Exposes process state statistics from /proc.
thread	Exposes thread statistics from /proc.
thread_wchan	Exposes thread_wchan from /proc.

Prometheus metric names

The following tables describe the Prometheus metrics exposed by each collector.

Process-level metrics

Prometheus Metric	Enabled by	Type	Description
process_context_switches_total	context_switches	counter	Context switches (voluntary and non-voluntary).
process_cpu_seconds_total	cpu	counter	CPU usage in seconds (user and system).
process_fd_ratio	fd	gauge	Ratio between open file descriptors and max limit.
process_major_page_faults_total	memory	counter	Major page faults.
process_memory_bytes	memory	gauge	Memory in use (virtual memory and RSS).
process_minor_page_faults_total	memory	counter	Minor page faults.
process_num_threads	thread	gauge	Number of threads.
process_open_filedesc	fd	gauge	Number of open file descriptors.
process_read_bytes_total	io	counter	Number of bytes read.
process_start_time_seconds	start_time	gauge	Start time in seconds since 1970/01/01.
process_states	state	gauge	Process state (R, S, D, Z, T, or I).
process_write_bytes_total	io	counter	Number of bytes written.

Thread-level metrics

Prometheus Metric	Enabled by	Type	Description
process_thread_context_switches_total	thread	counter	Thread context switches (voluntary and non-voluntary).
process_thread_cpu_seconds_total	thread	counter	Thread CPU usage in seconds (user and system).
process_thread_io_bytes_total	thread	counter	Thread I/O bytes (read and write).
process_thread_major_page_faults_total	thread	counter	Thread major page faults.
process_thread_minor_page_faults_total	thread	counter	Thread minor page faults.
process_thread_wchan	thread_wchan	gauge	Number of threads waiting on each wchan.

Threading

This input always runs in its own thread.

Getting started

Configuration file

In the following configuration file, the input plugin process_exporter_metrics collects metrics every 2 seconds and exposes them through the Prometheus Exporter output plugin on HTTP/TCP port 2021.

fluent-bit.yaml

# Process Exporter Metrics + Prometheus Exporter
# -------------------------------------------
# The following example collect host metrics on Linux and expose
# them through a Prometheus HTTP endpoint.
#
# After starting the service try it with:
#
# $ curl http://127.0.0.1:2021/metrics
#
service:
  flush: 1
  log_level: info

pipeline:
  inputs:
    - name: process_exporter_metrics
      tag:  process_metrics
      scrape_interval: 2

  outputs:
    - name: prometheus_exporter
      match: process_metrics
      host: 0.0.0.0
      port: 2021

fluent-bit.conf

You can see the metrics by using curl:

curl http://127.0.0.1:2021/metrics

Container to collect host metrics

When deploying Fluent Bit in a container you will need to specify additional settings to ensure that Fluent Bit has access to the process details.

The following docker command deploys Fluent Bit with a specific mount path for procfs and settings enabled to ensure that Fluent Bit can collect from the host. These are then exposed over port 2021.

docker run -ti -v /proc:/host/proc:ro \
            -p 2021:2021        \
            fluent/fluent-bit:2.2 \
            /fluent-bit/bin/fluent-bit \
            -i process_exporter_metrics \
            -p path.procfs=/host/proc  \
            -o prometheus_exporter \
            -f 1

Enhancement requests

Development prioritises a subset of the available collectors in the third party implementation of Prometheus Process Exporter. To request others, open a GitHub issue by using the following template:

• in_process_exporter_metrics