robustperception/demo_prometheus_ansible

Ansible config for live Prometheus demo

demo_prometheus_ansible's Issues

example config regarding memory chunks and local retention time

I had a look through the main.yml file and had some questions with a few configuration aspects.

 # Only 1GB of RAM
 DAEMON_ARGS="$DAEMON_ARGS -storage.local.memory-chunks=51200"
 DAEMON_ARGS="$DAEMON_ARGS -storage.local.max-chunks-to-persist=51200"

What is the relationship between memory chunks and RAM? Is there a fixed size to a chunk?

# Keep 6 months of data
DAEMON_ARGS="$DAEMON_ARGS -storage.local.retention=744h"

744h equals 31 days -- how does this result in keeping 6 months of data?

RobustPerception prometheus packages

Hi,

I'm not sure if this is the right place to report repository issues, but anyway here goes.
I'm trying to use your repo and packages with ubuntu Xenial and found out that the names of packages differ. Would you consider changing packages names to match the names of official ubuntu packages eg.:
node-exporter -> prometheus-node-exporter
pushgateway -> prometheus-pushgateway
alertmanager -> prometheus-alertmanager

Currently, when your repo is enabled on xenial, the search for prometheus results in:

ring01:~# apt-cache search prometheus
cadvisor - analyze resource usage and performance characteristics of running containers
golang-github-google-cadvisor-dev - analyze resource usage and performance of running containers
golang-github-prometheus-common-dev - Common libraries for Prometheus components
golang-github-prometheus-log-dev - Logging library for Prometheus's Go-based components
golang-prometheus-client-dev - Prometheus instrumentation library for Go applications
prometheus-cli - Prometheus command line interface
prometheus-node-exporter - Prometheus exporter for machine metrics
prometheus-pushgateway - Prometheus exporter for ephemereal jobs
python-prometheus-client - Python 2 client for the Prometheus monitoring system
python3-prometheus-client - Python 3 client for the Prometheus monitoring system
prometheus-mysqld-exporter - Prometheus exporter for MySQL server.
alertmanager - Alertmanager for Prometheus monitoring system
node-exporter - Prometheus exporter agent for machine metrics
prometheus - Prometheus monitoring system and time series database
pushgateway - Pushgateway for Prometheus monitoring system

Possible issue with failure cases in the cronjob here?

*/5 * * * * root du -sb /var/log /var/cache/apt /var/lib/prometheus | sed -ne 's/^\([0-9]\+\)\t\(.*\)$/node_directory_size_bytes{directory="\2"} \1/p' > /var/lib/node_exporter/textfile_collector/directory_size.prom.$$ && mv /var/lib/node_exporter/textfile_collector/directory_size.prom.$$ /var/lib/node_exporter/textfile_collector/directory_size.prom

demo_prometheus_ansible/roles/prometheus/tasks/main.yml

Line 97 in e063538

*/5 * * * * root du -sb /var/log /var/cache/apt /var/lib/prometheus | sed -ne 's/^\([0-9]\+\)\t\(.*\)$/node_directory_size_bytes{directory="\2"} \1/p' > /var/lib/node_exporter/textfile_collector/directory_size.prom.$$ && mv /var/lib/node_exporter/textfile_collector/directory_size.prom.$$ /var/lib/node_exporter/textfile_collector/directory_size.prom

@brian-brazil One issue here is that the '&&' will only fail if the 'sed' command fails, and not if the 'du' command fails... e.g.

prometheus1 ~ # du -sb /var/lib/prometheus | sed -ne 's/^\([0-9]\+\)\t\(.*\)$/node_textfile_directory_size_bytes{directory="\2"} \1/p' && echo WORKED
node_textfile_directory_size_bytes{directory="/var/lib/prometheus"} 80169933421
WORKED

prometheus1 ~ # du -sb /srv/prometheus_missing_dir | sed -ne 's/^\([0-9]\+\)\t\(.*\)$/node_textfile_directory_size_bytes{directory="\2"} \1/p' && echo WORKED
du: cannot access '/var/lib/prometheus_missing_dir': No such file or directory
WORKED  ## this worked because the 'sed' command succeeded, even though the 'du' command failed

You might be able to resolve this by wrapping the commands into braces to invoke them together into a sub-shell? But, In cases like this, I think it would be better to recommend using a small 'wrapper' shell script for cron to execute, rather then trying to squeeze all the logic into one cronjob line?

If that sounds reasonable, I could work on a PR to update this playbook to correct this issue.

Add support for https in Nginx proxy

Hi,

For now, Prometheus console is only available in http, on port 9090.

I would like to use your Prometheus instance as a backend for the demonstrations of 2 JS library I'm working on:

• https://github.com/samber/prometheus-query-js
• https://github.com/samber/chartjs-plugin-datasource-prometheus

Github pages are hosted with HTTPS URLs. Browser are not allowed to send requests to insecure HTTP endpoint.

Are you willing to add TLS to this nginx proxy ? Do you have a robustperception.io wildcard certificate ? Do you prefer LetsEncrypt ?