Have a read of the first post https://infotechmentor.com/2026/05/01/how-i-break-down-a-system-diagram-part-1/, this will provide background to part 2 of this post.
Part 2 of this post will provide specifics on the overall understanding of the system architecture diagrams. I will go into break down each of the diagrams below and provide examples of the types of questions around the systems.
The examples will be in italics and coloured format, which will differentiate it from the original parts from the first post.
Have a view of the following diagrams below. These are typical high level system/service architecture diagrams which tech companies/departments will show their IT users. These are shown to the IT users to grasp and provide an understanding the flow of systems.



Looking at these diagrams for the first time as a junior employee or even an experienced employee can still raise the same feelings. Uncertainty, confusion, am’ I getting this right, what am’ I missing etc etc.
You do need to know some basis of systems knowledge.
- Front end
- Back end
- Middlware
- API’s
- Hosts/servers
- Connectivity / handshake
- HTTP
- HTTPS
- FTP
- Source and destinations
- Types of scripts e.g. .js / .java / .py
- 3rd party system and vendors
Below are some questions that will aid you in having some general understanding
For these example I will be describing an internal email system for clients and a content management system (cms) for the marketing team. Further to add the terminology and explanations are fictitious but still emphasise the main idea of explaining parts of the system. Further to add the users mentioned on this post are fictitious
- What is the purpose of this system and what is is meant to fulfill
- What happens if this system fails? > Is the client able to access the website, cms, backend tool to perform their work. Like reporting, emailing, pressing buttons on the cms etc
- Who is affected > Business operations, tech operations, help desk, back office. Jack and Jill Doe? What are their contact emails and numbers
- What needs to be done to restore / repair the system > Show me the procedure and steps to restart the gateway, middleware, and backend services
- Where are the current weaknesses or soft spots in the system > We have this looping behaviour in the email system where it sends emails twice. We currently have our email provider looking at the issue. There is latency in delivering the reporting api message between our accounting team and the Adobe cms. We’re working with Adobe and the IT ops team to investigate and get to the root cause
- What part of the system always breaks and needs human intervention > As discussed, our email system sends emails twice, so we need to jump into the e-mail server and delete the second email before the user sees both emails. As for the latency we have an IT ops engineer monitoring a dashboard that will show us the spike in latency. We don’t have a work around at this stage for this issue. We’ve inform the clients both the sender and receiver to wait a few extra seconds
- What does the IT user do specifically > The users affected by the email issue are our internal office users which need to send daily transaction reports to our clients paying for our software service. If the internal office users see two emails they will immediately forward the client both emails, billing the client twice. That is going to be terrible. As for the API latency, the users mainly affected are the marketing team that need to put together promos and discounts into the Adobe cms. If they are waiting there for extra seconds, it you multiply that by days, the seconds becomes minutes. Time that can be allocated to other duties and tasks.
- Are there procedures and instructions to repair and restore > We have documented procedures in the jira system. We have logged tickets with the IT App Support Ops Helpdesk, so they are tracking each event. The procedures are in step by step numbered format (at this point you will be shown the documents and you will be given a dry run on how the procedure is executed).
- What happens if this system fails? > Is the client able to access the website, cms, backend tool to perform their work. Like reporting, emailing, pressing buttons on the cms etc
- Where and how does the client/customer enter
- Does the client enter via a URL / website link / button etc > The emailing system the office users access is through the desktop .exe called emailClient.exe, this is a launch able icon on their desktop, for MAC users they need to log into our windows jump box and then access the .exe file. As for the marketing team they can access the Adobe cms portal through the web url http://www.thisisnotarealadobelink.com where they will be able to access the promos and discounts
- What does the client do in this system
- E.g. do they enter inputs like numbers into a form, does the button perform some send action which gets relayed across to another user, is this a reporting system where the client clicks on some buttons to generate a document etc
- > The emailing system is needed so we can bill clients for their software usage. We use this billing system as well to track the types of activities of the clients. The Adobe cms is where our marketing teams put together promos using the design interfaces and widgets, along with putting together discount incentives through the discount api service offered on the Adobe service
- Is there a demo somewhere end to end where I can see how this system is meant to perform and behave
- In other words can someone show you how the system works, every step, all inputs and outputs, and everything that happens in between
- > We have a sandbox account that can be accessed that shows each of these system. We’ve made sure we have test environments to demonstrate to new staff and test new features. We will need to grant you access to the sandbox first. The functionality is exactly the same as we described earlier.
- What are 3rd parties or external vendors are we reliant on for our system to work > The email system emailClient.exe is built in-house by our developers. The code is store locally in our GitHub onprem system, so we’re sustainable on our own with this. As for the Adobe cms system we do have a ops support team that looks after the level one aspect of the system (i.e. restart, check logs, servers, app containers etc), but if they are not able to diagnose or identify root cause of an issue they usually need to log a ticket with Adobe. Adobe turn around time is usually 24 hours or less
- E.g. does this system connect to an offsite database, a cloud backup system, another website or source that is not managed by our company etc > the emailClient.exe is hosted locally, you can access all the servers and running app code on the developer portal. You will need an ide and these secret keys to be imported on our machine. The Adobe cms is in the cloud. We can provide the username, password and url
- What are these 3rd parties and vendors
- Who and how do we contact these external parties > to reach the operations team supporting the clientEmail.exe system, drop them an email at ops@thisisourcompany.com.au. Their number is 1300 123 456, but drop them an email ticket first. The Adobe provider, we have a client manager, and we reach them on thisisourcompany.com.au@adobe.com. Don’t go through their public portal otherwise they will think we’re a new client, we’re an existing client. I usually speak to a Joe or Jan on the other side of that email. After they establish contact via email they usually provide a teams room number where we can arrange meetings etc
- What do they do, what do they provide to us > the devs provide us the backend support for the clientEmail.exe system, they know the ins and outs out the app. The only thing they might be light on is the networking side. We do have a few unique routing for this app. We do have a networking team that can also assist. If there is a failure in executing or running of the app or specific function get the devs to take a look. As for the Adobe system, we have a 24/7 support model. Most issues they are able to help us on. We’ve had the odd case where they’ve needed to escalate higher up on their side, otherwise they’ve been quite responsive. Here are some tickets we’ve logged previously on previous issues
- Is there something in this system that is hidden that I need to know about(?)
- Get the employer to be honest, challenge their thinking and question their knowledge
- This isn’t a gripe question, but if you are to support this system you want to know all aspects to the system
- > Our support model here is pretty transparent we’re not hiding or being non transparent on how we manage our systems. I will admit due to the amount of layoffs knowledge has been a bit more scarce. So if you don’t get the answer straight away that’s probably why. Just also be wary of the networking team, they tend to cop a bit. There is a tendency to blame networks, do the due diligence first in troubleshooting (i.e. provide src + dest ip’s and the urls you’re researching), before escalating to them. Otherwise you have access to our documentation directory in jira. Feel free to have a read of the doco’s there
- Just on a side note, if there are hidden aspects to the system (either for non malicious or malicious reasons (you never know)), just take a note on it for now. Maybe at a later time after some experience and knowledge building you can raise it later again with your employer > For obvious reasons you can’t get access to the financials of the company, that’s only for our executive board and leadership teams to have access too. If you see a lock sign on the some of the folders and files that’s probably why.
- Get the employer to be honest, challenge their thinking and question their knowledge
Summary:
- Begin by asking some basic questions of where and what
- Don’t ask why. You’re there to support what’s in front of you
- There will always be discussions later on optimisation and improvement
- > Above is examples on how to communicate your questions to your staff and leaders. I use technical terms, words and examples on how you should communicate in the IT workplace. It will make you look very competent
- Don’t ask why. You’re there to support what’s in front of you
- Document the weaknesses and soft parts of the system which are known to vulnerabilities and breakages
- Get steps and instructions on how to resolve these weaknesses
- > Get your team lead, manager and other staff to explain in detail specifics of what you’re supporting. This goes for IT ops engineers, devs, devops, networking engineers, team leads etc. Not just for IT users in the operations and level 1+2 space