5

As wiki says:

Monitoring can be performed by continuously measuring certain parameters by using a medical monitor (for example, by continuously measuring vital signs by a bedside monitor), and/or by repeatedly performing medical tests (such as blood glucose monitoring with a glucose meter in people with diabetes mellitus).

And in well equiped medical offices we can see patient monitors at every bed. I guess there is also some alert system connected to this stuff. But monitor is always on.

enter image description here


In software system I am working with there are several security bridges for RDP connection, so these all bring lots of ancertainty to the behaviour of a system.

And I am controlling it's availability by some alerts, but sometimes alerts are not fatal, and it looks like no possibility to destinguish fatal from common. Be sure I am trying to controll these differences, but sometimes we have new error and testing system can not handle it properly (it is tested by , because it is RDP connection). So I prefer to follow next steps:

  1. Recieve alert message
  2. Check (look at graphs) whether it is on both testing machines (this part could be automated)
  3. If it is on both, check testing machine
  4. Futher actions

So my question is

Is it Ok to have "patient monitors" at the department wall or infront of someone who receives alerts or at developers department or...?

UPDATE:

Now I see this monitoring board (big screen) like this (1. free JVM memory; 2 & 3 Apps availability), not just list of alerts (graphs are created with Zabbix):

enter image description here

2
  • 2
    My employer does a lot of this.
    – user246
    Mar 29, 2017 at 11:25
  • 4
    closing mafia again? Yes, it is complicated question - life is complicated. Mar 29, 2017 at 19:34

3 Answers 3

7

enter image description hereIs it common/possible/recommended to have "patient monitors" at the department wall or infront of someone who receives alerts or at developers department or...?

Oh yes. You'll see this most intensively used in devops where they are often 'walls of monitors' with big screens showing the status and activity in the various systems being monitored.

I've seen this extended into application development areas. I've even heard the term 'bvc' as in 'how comes you guys don't have a BVC. BVC means Big Visible Chart and most commonly refer to a large monitor in the development area for each group.

A typical setup that I've seen is to have a large monitor with a split screen, displaying:

  • current build(s) in the ci server
  • current ticket board and assignments, e.g. Jira

I've also seen a monitor continually running UI automation to ensure the production site is working. A downside is that the flashing screen (from page transitions) can be distracting.

The image below shows the sort of setup that devops might have.
For teams of developers it is more likely to be one screen, possibly with split window views.

enter image description here

side note: some folks use a paper board and stickies to intentionally avoid needing the technology, cost and physically demands of a big screen. The main downside to this is that remote or work-from-home-day workers can't see it, interact with it, etc.

6
  • 2
    I'd like to double upvote for this photo :) Mar 29, 2017 at 15:23
  • Your last sentence reminded me of a previous job where we did use a big whiteboard and stickies, but then had webcams streaming live images of said board to other sites and remote workers. It worked pretty well.
    – Vality
    Mar 29, 2017 at 17:10
  • 1
    Yes. Excellent solution @Vality! Mar 29, 2017 at 17:13
  • ok @IvanGerasimenko I've added another photo at the start, check that one out! Mar 30, 2017 at 1:56
  • @MichaelDurrant, that one is not about application/system monitoring & software testing Mar 30, 2017 at 7:00
1

Monitoring is a key part in the quality of software product. Mainly after deployment, but operations and maintenance is part of the SDLC.

Who is responsible for following up which monitoring alerts should be defined per company. In the beginning this could be the development team. If the team and the application is small. As the user-base becomes bigger, teams like support or devops might be more appropriate. Just experiment with what makes the most logical sense now and continuously improve the monitoring (follow up) process.

We actively visualize all our test automation and monitoring. We created a radiator screen to show any issues, but also we use slack/mail integration to update relevant teams in our business when data is not being processed for example.

Visualizing always helps when noone is taking action. We went from ignoring the monitoring systems totally to responding and handling within acceptable time-frames. Just by displaying the status in the hallway, the mentality changed with it, not by mandating change.

enter image description here

1

Excellent way to create your own custom web-accessible monitors is https://grafana.com/ - you can have graphs, dashboards etc generated from live monitoring data.

Grafana (and nagios) is cornerstone of our (in our company) way to visualize our system, and gain insight what might be causing problems.

Grafana converts raw data into actionable information.

Disclosure: I have no relation to Grafana project, only a happy user.

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.