dashboard options

We'll first want to run a dashboard on our local system, using a light webserver and PHP server, before attempting it over a LAN or WAN. Step one is to inventory all our databases (including browser's IndexDB), logs, mail, and daemons (units/services) on our systems prior to attempting the dashboard. That's because a dashboard will add to what's already present. We need a system baseline. Because for example, the simplest MTA's add a database of some sort. SMTP4dev adds an SQLite.db in its directory, same with notmuch (or maybe an XML datatabase). If we go so far as Postfix, it requires a full relational database. So we need a pre-dashboard inventory of databases, logs, and mail, written on a static HTML page.

why dashboard

We may want to monitor certain applications or system parameters, financial counts, student attendance, anything. If we start with our own system as a model there are 4 things which regularly change: logs, timers/cronjobs, system values (temp, ram usage, hdd usage, net throughput, etc), and application information. We might even want to craft a database to store some values. Postgres is best, but we can get general theory from MySQL models.

MySQL table basics (14:26) Engineer Man, 2020. Great overview for the time invested.

Prior to a real-time dashboard, a slower process of daily emails with a summary are a good start, even if we just mail them to localhost disconnected from Web.

Once we can update a daily localhost email, we can attempt to expand that to internet email. Or/and, we can add a light local webserver + PHP. We need this to have dynamic webpages; opening an HTML webpage from a file directory is "load once", not updatable.

configuration models

Security camera configurations were created around updating pages, and these configurations might be adaptable to system monitoring. More than security models, the DevOps/BI models seem glean-able. These servers might be, eg Fluentd, Prometheus, and Grafana. Production server software, but localhost possibilities. Most are in the AUR. Prometheus is sort of a DAQ for software -- with email alerts possible -- and then Grafana is typically the display for it. But neither require the other. Grafana can display from any data source, or take its info from Prometheus. For Postgres built DB's, TimescaleDB has a lot of videos that might apply. We might even be able to modify a Moodle setup, now that we can upload quizzes using the Aiken model.

services

We can attempt several daemons on a local machine, and see which ones are too resource intensive. Also timer scripts to execute those services only as needed, stopping after.

local system dashboard

logs and timers

Nearly all logs are inside /var/log, but we need to evaluate our system carefully at least once for all relevant log locations. Some logs are ASCII, others are binaries that require an application or command to obtain their info. Once tallied, systemd timers and scripts are the simplest, with possible output via postfix. If we then add a webserver and PHP, we could run a systemd timer script every hour on logs and which updates a localhost webpage. To see running timers...

# systemctl list-timers

Fluentd is a log aggregator available in the AUR, but may also need a DB to write to.

Timer files (18:16) Penguin Propaganda, 2021. This guy makes one to run a script. However, what about running a program and then shutting down the service after? 10:00 restarting systemctl daemons and timers.
Systemd start files. (7:52) Engineer Man, 2020. How to make unit files, which we would need anyway for a timer file.
Dealing with system logs (10:00) Linux Learning, 2018. Mostly about /var/log. Explains some are ascii others binary.

email inside our system (mua + mta)

Link: localhost email setup - scroll down to email portion.

For a graphical MUA, I used (in Chromium) an extension called Open Email Client. Some configuration information is provided by the developer.

various

ISP Monitoring (8:04) Jeff Geerling, 2021. Jeff (St. Louis) describes typical frustrations and getting information on power and internet usage. Shelly plugs ($15 ebay) are one answer. However there are excursions into several dashboard options.
How Prometheus Works (21:30) Techworld with Nana, 2020. Why and how Prometheus is used.
Grafana Seminar (1:02:50) TimescaleDB, 2020. Avthar Sewrathan (S.Africa) Demonstation and some of his use cases
Grafana Seminar for DevOps (1:01:59) Edureka!, 2021. Grafana half of a Edureka seminar on DevOps with Prometheus and Grafana. Thorough description including how to create a systemd service file to run it locally.
Prometheus and Grafana (54:06) DevOpsLearnEasy, 2020. Adam provides description of a server deployment of Prometheus and Grafana. This guy even shows the setup of a VM on AWS. He seems confused but we learn, same as strike football penetrates the veneer.
Grafana Seminar (1:02:50) TimescaleDB, 2020. Avthar Sewrathan (S.Africa) Demonstation and some of his use cases
Grafana Seminar for DevOps (1:01:59) Edureka!, 2021. Grafana half of a Edureka seminar on DevOps with Prometheus and Grafana. Thorough description including how to create a systemd service file to run it locally.