2

I have a system that I am monitoring. The system consists of a master server with multiple smaller servers underneath it. They all communicate with eachother, perform different tasks and so on. Sometimes things goes wrong, which creates an alarm in the server. Since the system could consist of up to 200 servers, it's very tiresome to browse all the servers for potential problems and alarms.

So I want to create a tool that can look at the servers from the outside, check if there is a problem or not. I want to visualize this information in a dashboard like web tool, which could be something like a todo-list for the users monitoring the system. I want this "list" to be updated with potential system problems, giving the user a hint of why an item was added and I want the users to be able to assign themselves one of those tasks. Then they'll go into the system, investigating, fixing the problem, and when done somehow ticking of the item from the list.

What would be a good way to visualize this? Is it just a simple list of the issues or is more detailed cards a better solution? When thinking about this one, it feels as if it must already exist similar solutions out there, but I can't really find any good examples.

Improve this question
5
  • This reminds me of the "Suggested Edits" review system here in Stack Exchange.
    – Alvaro
    Commented Jan 29, 2017 at 19:01
  • @Alvaro: Can you elaborate on how that system works, or point me in the direction where I can learn more?
    – Michael
    Commented Jan 29, 2017 at 19:08
  • Sure! It is not the same, but it might give you some idea for your case. In this question there is information and links with screenshots.
    – Alvaro
    Commented Jan 29, 2017 at 19:15
  • About your question, how to represent the information depends on the kind and the amount of information you have/want to give.
    – Alvaro
    Commented Jan 29, 2017 at 19:31
  • @Alvaro: Thanks for the extended information. I like the idea of a queue like list, that will add things as there is room and users can execute items one by one. The amount of information would probably consist of a time stamp when the error occurred, a one sentence description of the error, a one word category, number of occurrences and the name of the server having the problem.
    – Michael
    Commented Jan 29, 2017 at 19:45

1 Answer 1

Reset to default
1

I have no experience with it but these are some thoughts, too long for a comment, in case they help:

What I understand you have is a list of elements with several fields. A good option could be to order the elements depending on the field (by date, by category, by number of occurrences, etc.). A data table with sortable columns could work.

It might be relevant to find any possible relation between the problems, either in the same server or in different ones. Maybe there is an specific problem that is causing others. To find these relations some kind of analytics system where you can inspect variables in a timeline, in a graphic with data usage...

I would go to the source of the process. Analyze what are the possible problems, how and why do they happen and then build the tool trying to represent anything relevant from it. Indicate the problem by itself and within the context.

Improve this answer
1
  • Thanks! From your answer I can see that you understand my problem. I think that the relations between problems is a very interesting concept, because that is what goes on behind the scenes in people's minds today when manually browsing the system for errors. They need support in that process, so they can focus on actual problems and ignore checking things that have not failed or are errors that comes of something else failing.
    – Michael
    Commented Jan 29, 2017 at 21:03

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.