Lance Knight, ConnectAll | DevOps Experience 2022

Hello, you know there's a lot of challenges today with devops being everywhere, you know data is everywhere everywhere with that data. And the challenges is how do I get that information that was measurements in the right place so that I can create metrics analytics and make business decisions on one of the interesting things is that organizations today are really business driven or data driven. Like we think they are right a lot of organizations today barely even get to their data or know where it is to use it to create measurements.

And it's a big problem today. And even for devops people. It's a problem.

You know, hi, I'm Lance Knight. I am president. CEO of company called connect all I've been involved in it software development delivery for decades now specifically helping organizations to be more efficient with software development and delivery and focusing mostly on how you can use lean manufacturing principles.

I learned as a child to as a child. I say as a younger man to be more efficient. You know the thing is is that most it operations people devops people are just getting just getting saturated with requests today for all types of analytics and metrics from flow metrics devops metrics, you know, Dora metrics Improvement metrics, there's portfolio metrics agile metrics all of these types of knowledge and all of this data exists all over the place and they're struggling trying to figure out where do I get that information from?

You know Bob here, I'm sorry Dave here. Yeah, he's a devops manager and on a day-to-day basis. He's inundated with requests about how do I make my team more important?

You know, that's what's going on. How do I Get more metrics, for example, the Dora metrics can be confusing but there's also other types of analysis and metrics that are people are looking for and asking Bob from and his boss always comes to him says, how do I make the team more efficient? How do I measure that I am where do I get that information from is what Bob is asking or Dave here is asking in the back of his head.

So it's truly a complicated problem and as teams grow and things happen. You know and it's this thing it's like where is all this data? So if you look across larger organizations it even gets more complicated.

So how do you understand that? How do you get to a point where you have a method or a process where you can uncover that? In here's a bigger problem that as well when you start looking for that data.

It's in the cloud. It's in somebody's tools. It's on a spreadsheets and somebody's email and this becomes really challenging to know where to find this information so I can centralize it and do something.

It's existing in these silos and but unlike with other types of things that you use principles like value stream management for it could be anywhere. You don't see it. It's in the ethernet it sits in the computer.

It's on a server somewhere in Iraq in VM in Azure or something. So it's hard to even know where the data is because you can't just walk out and say it. And to go along with all the things that I just talked about there has been an explosion of tools.

There are thousands of tools just in specific industries that are around software development delivery. There are so many solutions out there today that have different niches and data exists. In all of these.

So how do you know when or where how you're using them or what they're doing for your organization? And how do you determine where the data is what data is important? Even more than that.

When when do I capture that information in order to pursue produce that metric and it becomes quite a problem. So a lot of organizations have multiple deployment tools multiple build tools. A lot of organizations have multiple agile tools and multiple portfolio tools.

And you know, I've seen this over the decades where you need to figure out how to get that information correlate it so that you compete data-driven. It's just complication is driving that first problem that I laid out of organizations not being able to get to be data driven is the data is everywhere where things are getting developed delivered and and put together. So my point of view that I wanted to give everybody today is I believe that you can use these principles that I've learned over the years to help you understand where the data is what tools are being used.

How work is happening. So you need to understand how work happens in order to know where things are going through to be efficient and I believe you can use systems thinking and value stream management principles to uncover where the information is where the measurements are and where that data is to be really efficient. In the first question, I always get asked is okay.

What is value stream management? I'm sorry. What is systems thinking and what can how can I use that?

Well, you know systems thinking is really that it's thinking about everything as a system a system is a set of things that you do in order to develop something or have something so you have inputs outputs and then things in the middle called elements. This is a it's very simple example of what I mean. And this is a clothes dryer everybody.

I hope everybody knows what a clothes dryer is and how you can use it to understand what systems are a clothes dryer is a system. It's enclosed system that does a specific thing and there's the black box or black blocker Black Box representation of that system where you know, you put in wet clothes you add cold air and electricity. It goes through its phases of transition and outcomes.

Dry clothes and that's your cycle time for that. But more happens within this system that can affect that can have impact on that. It's more to it than that, but when you use it, this is all you understand but inside there.

Is more than that there are other things for example, there are elements in there elements that are the heater the drum the electric motor the fan all of these things work together to Tumble and dry those clothes that you can take them out. Those different elements have different things just like just like in software development delivery. There are elements that you use to build.

Run tests and deploy against so you know when you think about where all this data is it exists Within These elements of your software development delivery. And they have feedback loops and there's communication and Synergy happening between those so all of this happens in the right way and there are things that are protective like, you know, if the stove pipe and air can't flow at the backside can't come out. It'll stop the dryer because it's clogged up and and so on those are feedback.

Loops are happening. Um, but when you take a look at these things you realize and this is you know, getting back to the theme of where where I can pull data information is being calculated at each one of those that information is used for a feedback loop that information can then be used to put into a place where I can draw and create metrics and draw analytics from it to say how many times at this failed and use that information or to be more effective with drying clothes and looking for bottlenecks. So that's a system and that's just some thinking and I think you know, I often think of that as kind of the precursor to you know, value stream management and value streams.

And that's where we get to talking about value stream management principles. Now this is a value stream. I think we've all seen these over the years recently the it's a big thing recently, but then been around for decades and simply put a value stream is all the steps it takes To take requests from the customer both value added and non-value added in order to deliver products and services.

So you can see this is a pretty much a standard value stream. We've seen them all before. In order to understand where to get data though and where to grab information for the metrics that you've been asked for.

I think you need to start to use this one principle called value stream mapping and take a little bit further. This is an example of a value stream here. Now.

This is a like this is a three-tiered architecture. I learned about a long time ago. This is a governance model that you can look at.

You can see that it shows both ends of this and every organization this is different, but in order to understand where to get data and where to find information have to go through an exercise like this so you can see how things are flowing. So for example, there is an epic story in that came from the business as a request and it's here an intake where somebody's looking at it understanding and it's going through it's processes. Then all of a sudden it moves over to improval.

Information was just set up there a measurement just happened in a solution where I may need to capture that so I can provide information or provide analytics around that. So the breakdown of work the flow of work is all part of that as well as I move forward to the next page. You see I've moved to the design phase of my portfolio level.

And that actually triggered creating a new artifact in the development flow user story. Another user story is going to go through its moments and data was captured because right then that epic triggered somehow for a user story to be created. I want to catch capture that I want to be able to put analytics together with it.

And as I'm moving forward to keep some uncovering yourself, you can see how things are moving through your value stream here with this and the reason why I'm showing this I've used this diagram quite a bit. It really does show How information is being? Correlated captured and happening in your value stream.

And you have to create some type of map to show all of how this happens. That way you can start to look into there and find out where can I get information things like cycle time and lead time. I'll talk a little more about that as I go on you can see it's still moving forward and in things are moving forward package was created here.

It's been built tested release now, it's deployed operation has that particular package now other microservice or whatever and they're still information being captured. It is existing all through this method and process. And all the way over to you know defects to uncovered.

Now once you have this now I can say all right. Now I understand what's happening and start to figure out where to capture data from here and what to do with that data information. Now I can do and now I can overlap all the tools the solutions the things that I are used to do this this information exists in rally here where I'm doing the whole portfolio stuff that communicates with what the development's doing in in jira, and I start to figure out where correlate this information comes all the way down to where builds are happening in Jenkins and tests are happening and selenium and I can figure out where this data is based on the metric that I need to produce and start to understand because I know what the system is.

I've mapped it out and I can go find information in here and start to be more effective rather than going. Oh my God, where is all this data? I have a blueprint to go find it with through my organization.

You can drill it down further in each one of those flows. So here, you know, I could start to calculate cycle time through this flow and I can actually bring that together with the other flows because all of this is is now visible to me where I can actually see how things are happening where in the past I couldn't even see this and I can see where this would happen in different tools. So a lot of data is being captured about my backlog in my Jared tickets in jira as well tests are happening in my microphocus testing suite and now I can say are here's where I can start to get things like defect density and all those things exist and I have cycle time there and with operations too, you know, here's all the devops things that are are there as well.

I can look all the way out to where you know, I might have a list of my operational tools that are running scans and looking for vulnerabilities or selenium vulnerability tests in the middle and I can try to understand what's my meantime to repair because I know that that information exists in these other Solutions and I can start to figure out where I can get and correlate this data across my whole value stream here by understanding this principles and using it to map it out. And I often talk about these ears are site pox. But you can go all the way down to drill into the next step of those tools and say Here's those processes and each one of those processes data and information is being created.

So what I'm trying to say is you you can you can look at this from this way. It'll let you know where the information you're going to need is being created in the methods and processes of developing software. Without this you're kind of guessing at what's really going on.

Now, you have a clear understanding you're able to see where it's happening what the different things are going on within your software development system or value stream. And be able to see where I can capture data, but now I understand that now, let's go back to the other problem about where metrics are everywhere. Everybody's coming to a day to say I want this metric.

I want the flow metrics door metrics and you just have to really be able to step back and say what is what's important to the organization? What do I really need to capture in order to be effective? we believe in this thing called gold question metric and it's really simple goal question metric is What's the goal I have what I want to try to understand and what's the impact that I can get from understanding that goal?

Bottles are great bottleneck metrics are a great example. I can understand where things are being slowing me down inside my value stream and I can figure out how I can improve flow. But until I capture that and understand that what's the question?

How do I improve flow? And the last part is what metric do I want to look at? Maybe that would be cycle time or whatever those metrics are that people asking you for we start to look at them through these lenses.

So now that I know where the data and I can prioritize what metrics are really important. I can be more effective as a devops person. And again, it's it's quite complicated because devops is everywhere and tools are everywhere to work with and the devops mindset of feedback loops and all of that how many errors came back and all of those types of things so these things kind of help you out.

In these metrics then become relevant to where you can start to say, here's where I can take action on those. I know where the information is. I know why that's gonna be relevant to me.

And I know how I give that information in this visualization of a metric in the right way to what customer So the big question is what do I end up with? Well, you can end up with something like this. This is a data flow chart, you can go and create these there are different mapping solutions that you can go and use to to lay these things out.

We have one of our own Insight connect all that we use this for and then let's you tell the story of where you're going to get that data how it all connects and so on and you can create these pretty efficiency pretty pretty easily. Once you understand the flow of how things are working in software development delivery. And your metrics then are a lot easier to produce as well and you can say okay and I know how exactly how to go get that data where it is how to connect to it how to deliver that metric so that she could be more efficient and what you're doing today and remove costs and remove waste and look for removing the non value added activities in the organization by being data driven.

The question is how do you get started first learn about systems thinking learn about value stream management principles not about the ice cream Management Solutions. Those are nice, right we offer one, but it's more about understanding this age old value stream management principles learn lean understand lean and use that as part of your I'd say part of your your weapons of helping organizations to be more efficient and making your own jobs easier as well. Create a flow map in understand how to use gqm.

To help organizations understand what metrics are really important to you. Thanks for listening to me today. And I really appreciate you taking the time to hear me talk about how you could be more efficient with software development delivery and how you can find data to produce those metrics when devops is everywhere.