Azure DevOps & Power BI – Ep. 262

Treating BI artifacts like “files you click” is how teams end up with mystery changes, broken refreshes, and late-night copy/paste deployments.

In Ep. 262, Mike, Tommy, and Seth dig into Microsoft’s guidance on Power BI Projects (PBIP) and Azure DevOps build pipelines—and why Git-based workflows (PRs, branch protections, and automated checks) are the difference between hoping a deployment works and knowing it’s reproducible.

They also discuss what’s changing in Power BI deployment pipelines (including support for up to 10 stages) and how to decide when you should lean on Git, when you should lean on deployment pipelines, and when “more stages” is just complexity without clarity.

News & Announcements

• Power BI Projects build pipelines — Microsoft’s walkthrough for wiring PBIP into Azure DevOps so you can build, validate, and promote Power BI items with repeatable steps.
• Suggest a podcast topic — Submit an idea or question for a future episode.
• PowerBI.tips Podcast — The full Explicit Measures back-catalog, with links to listen on your preferred platform.
• Power BI Theme Generator (Tips+) — Create theme JSON that standardizes colors, fonts, and visual styling across reports.

Main Discussion

If you want a clean release process, separate the concerns:

• Git is for change management (history, review, accountability).
• Pipelines are for repeatability (the same steps, every time).
• Environments are for risk reduction (catch issues before users do).

Here are the key takeaways from the conversation:

• Use PBIP to make Power BI content “source control friendly”: storing models/reports as files unlocks diffs, reviews, and consistent packaging.
• Start with a boring branching model: protect main, do work in feature branches, and merge via PR so changes are intentional and auditable.
• Automate the “did we break it?” checks: even lightweight validation in a build pipeline helps catch obvious problems before they hit test/prod.
• Decide what promotion means in your org: deployment pipelines can move content through dev/test/prod (and now more stages), but you still need named environments and ownership.
• Don’t confuse “more stages” with “better governance”: adding environments only helps if each stage has a purpose (validation, UAT, training, hotfix, etc.).
• Keep secrets out of the repo: parameters, credentials, and tenant-specific settings should be injected by the platform/process—never hard-coded into versioned files.
• Adopt incrementally: pick one workspace, wire up Git + a pipeline, document the rules, then scale once the workflow is stable.

Looking Forward

Pilot Git + PBIP with one workspace, then add a single automated build/validation step—once that’s stable, expand to a multi-stage promotion path that matches how your org actually releases.

Episode Transcript

0:29 good morning everyone welcome back to the explicit met podcast with Tommy Seth and and Mike good morning good morning gentlemen it is thday it’s Thursday welcome welcome to a Thursday makes for a makes for a discombobulated week where they have the excitement yesterday of no more school for the week really and and then I’m like Yay no more work for oh no got to go to work it’s very funny because two days before the weekend do your kids always ask oh hey we have off schol

0:59 always ask oh hey we have off schol School dad do you work today I’m like work every day every day I work yeah every day like when you’re an adult children you don’t understand how much you work to in order to have a house that has heat and food you can eat lights that you leave on yeah the lights that you this is why this is why dad works so hard because we have like it feels like everything nowadays is like everything costs money everything to do anything is like I was the other

1:30 to do anything is like I was the other day I was talking with with my family I was just like hey feel like everything costs money you want to go apple picking money you want to find something to do in the weekends or money you want to go to a park money like I’m like what the heck dude it was it was a while ago but I actually ran into like a park that actually had a toll thing in the front I was like I feel this is the nature trail everything there’s playground it’s like right over there yes yes

2:00 there yes yes let’s do a quick I’ll do a quick little teaser here and we can talk about more intro things so this episode on the podcast we’re going to go through a brand new article around Azure devops Azure devops and the integration between that and powerbi if you’re around fabric at all you’ll notice that there has been a very deep integration between Azure devops and git and powerbi fabric so ruy Romano who is the PM for the pbip format has has produced an amazing article talking through some deployment

2:30 article talking through some deployment patterns you can use in connecting that with Git so that’s going to be our main article for today I wanted to know if there were any other topical items that Tommy you or Seth maybe found across this week before we go into our main topic funny enough the the deployment pipelines in that devops was going to be an intro I was going to send it to you guys I think before it’s like what this actually may deserve to be its own I I I think we you would definitely we definitely agree with you Tommy on that one this definitely deserves to be its

3:01 one this definitely deserves to be its own there’s there’s actually two major developments that I think are that have been announced around deployment pipelines so there’s a deployment pipeline for fabric and this is using GitHub and devops or git and devops then there’s another release Lee Benjamin announced that there is now a deployment pipeline can go from two to 10 stages this was also released and I think this is more of a harbi premium feature not specifically a fabric feature so in powerb premium you have

3:32 feature so in powerb premium you have this concept of deployment pipelines when it was initially released you only had Dev test and prod you couldn’t make more you couldn’t make less that was all you had and so they have now revamped the I’m to Microsoft’s credit here the technology is the same whether you’re going to have three or four or eight environments all of them doesn’t really matter how many you have but now the deployment pipeline lets you have named a additional deployment of pipelines and you can have up to 10

4:03 pipelines and you can have up to 10 environments to move reports and content through 10 environments can you up to 10 could you could you imagine yeah that’d be a little bit crazy like the max I’ve ever seen is four five like an R& D Dev QA uat prod yeah I I can stand some of them I would have to imagine there’s a mean I would have to imagine there’s a lot more testing going on when you have something large like that but I don’t why I would need 10 chat I need some help here like like what is the max

4:34 help here like like what is the max level of environments what would the reason I I like outside of what I just labeled I I wouldn’t see unless you’re Crossing departments or something I don’t know yeah I I don’t know I I even have some or I can definitely understand the organization point of we really want like two right there’s for for smaller so the more or the more levels you have the more work it takes to to move through those different environments right so having

5:04 environments right so having some some business units right smaller teams might use less environments they don’t need as much right there’s not as much testing going on like you build something make something unique and then deploy it or or push it push it out anyways I love how the sentiment in chat is like who needs more than prod like PR’s not sufficient EXA where do you where do you test prod where’s where what are you doing where is it why mess around I got it in prod

5:34 it why mess around I got it in prod that’s actually that’s actually really funny because I think there’s a funny t-shirt that says I only work in PR like yeah I think I’ve seen that one before I I I would yeah anyway I’ll just move [Laughter] on on honestly even with the pro I I think for a lot of business intelligence analysts like for the most teams the just simple Dev and prod just the two workspaces are are going to be absolutely critical like Yeah from if we’re actually going

6:04 like Yeah from if we’re actually going to see anyone with 10 please do a user group on it do a session on that why you need 10 user groups or 10 deployment pipelines 10 user groups yeah 10 user groups which as unnecessary as 10 workspaces in a deployment pipeline I I think so in that one as well it seems like process for process sake maybe something along those lines maybe there’s just anyways be still my heart that is Tommy’s World Tommy maybe you should just maybe you should create

6:34 you should just maybe you should create the use case for 10 H I’ll ask Chad GPT well if if you I think in in my mindset especially when I’m talking about diplomate pipelines I’m thinking about each diplomate pipeline being its own like let’s let’s imagine a deployment pipeline being a full start and a finish it’s literally a linear line right you’re moving from one environment to another I could see a use

7:04 environment to another I could see a use case here around where you have Dev and test and you might have like so that the two environments where you’re doing the work you might have six other and and Enterprise AR is speaking to this a little bit right now so I’m thinking about it you could have for each department in your organization right in deployment pipelines you can selectively pick what’s being deployed into each workspace so maybe this is an idea of like let’s think about there’s a central workspace Devon test where all the data

7:35 workspace Devon test where all the data sets and all the reports go and and then you’re trying to serve those group of reports out to four or five business units so you have everything makes it to test you can see and test the data there and then instead of having a single endpoint which is where my mind went right I’m thinking oh it’s prod it’s a single end point like one to one can you can you separate out one stage and break it into oh so so imagine imagine you have in this test environment you have all the HR reporting and the financial

8:06 all the HR reporting and the financial reporting and all the Ops reporting right but then imagine you’re deploying from test into HR workspace prod into operations prod or into HR prod right so now you can have all your reports follow the same process Dev and test and then you can then say okay here’s all the reports I’m going to push into this particular topic an app that’s for that team so so maybe the idea here is there’s not one endpoint there’s like

8:37 is there’s not one endpoint there’s like five end points there’s a Dev and a test and then five prods so maybe that’s where people are thinking about this where they would say I can then now so because we always had this issue of one workspace one app one app maybe maybe we need more apps off of those workspace maybe there’s still a lot of Legacy I think this would be a great User Group Tommy we should build we should build worldwide importers and how to deploy 10 objects give us more

9:07 how to deploy 10 objects give us more exactly we need [Laughter] more like kicking a gift horse in a mouth right is that the phrase yes I think so I don’t I’ve actually I think I’ve heard that maybe a couple times so anyways that’s a really interesting topic in of itself but I think this also leads us very well into the topic for today which I’ll the link is in the description already and I’ll also drop the link into the chat window so the announcement made by ruie here’s the link for the chat the the announcement

9:37 link for the chat the the announcement made by ruy is talking about using build pipelines and how build pipelines and Azure devops can be used you can and the overview of the article is it’s more instructional nature right connect fabric workspaces to Azure devops create and run an Azure devops pipeline Define your branch policies and then create a pool request to then take who are branching and push it back into the main branch which would then deploy the reports essentially I really like this

10:07 reports essentially I really like this and I think we’re we’re starting to see more investment by Microsoft more artifacts are coming to the deployment pipelines or even this get thing you initially had reports and patting reports I think they’re going to add more artifacts I couldn’t see why you wouldn’t add the spark notebooks in there too I meant it seems like a no-brainer it’s like it’s just a file so Version Control that as well I really like where they’re going with this I think this is a really good integration any thoughts gentlemen as we as we walk through this

10:37 as we walk through this one let me ask this question Tommy or Seth do you guys currently use git to control your files and are you really looking Are you seriously looking at using git now that we have the pbip format I think we’re getting there I definitely use it or have some test things definitely not in production or anything for clients okay just because it is still preview and there’s still some I’ve noticed some bugs if

11:07 still some I’ve noticed some bugs if you’re using one Lake and creating a new data set off of that and then using a pbip with Git for whatever reason okay I’ve had a lot of I’ve had a lot of issues with that with the data source o odd odd items but honestly the idea of at least from the data set point of view is fine but it’s going to be it’s not going to be our normal deployment if if you building something in python or you’re building something where it’s much more code based

11:39 where it’s much more code based because whether you’re doing Dax power query and then the only other really code in there would be working with the report Json file you your in a sense path or review or your your testing can’t really be necessarily on the code itself it’s not like yeah you can view it in code with all the all the G Integrations but you’re probably not going through it maybe like like with again with looking at change history but not nearly like you would for

12:10 but not nearly like you would for editing that’s where tablet editor comes in that’s where other tools come in and I don’t think anyone’s going to be editing Dax even with the new timle file language unless you could tell me otherwise H I don’t I don’t think I don’t think D is the main issue for me so let me give you where I think the PB P format really is strong here right and it’s it’s getting better and I think it will get better so one thing that will make this a lot easier is if you think about how power Bay reports are built today there is a single there’s a data

12:41 today there is a single there’s a data set definition and then there’s a report definition when you have the data set you need all the data in the data set and this was been one of the main challenges with using anything with Git previously this is I think one of the main barriers to entry in talking git with powerbi was in order for you to check that PB file in some files get extremely large and you’re now checking in a multiple like let’s imagine you have a 300 megabyte file that’s not out of the out of the realm of possibility

13:11 of the out of the realm of possibility for powerbi files you could easily have one of those with a lot of data in it so the data portion of that file is the large portion the stuff that I really want to Version Control I could care to some degree I could care less about the actual data because the file has here’s where to connect to the data here’s how to go refresh it when you pull down the definition really what I care about is the couple megabytes of report Pages the definition of the data model what are the columns Nam like all

13:41 model what are the columns Nam like all the meta schema of the data right so the important things I want to Version Control I don’t really care about version controlling the data because we’ll just refresh the report and get it so to me the the really magic of this PBP format is it separates out the ABF file that stores the partitions of all the tables when you’re working in Des Des toop to me this is the realce the nice Secret Sauce here it’s since it’s separating that away from the the definition of the file that’s huge because now all of my files that I

14:11 because now all of my files that I wanted to Version Control are now separate this being said there is a lot of other things that go along with this pbip format for example a lot of the report. Json files it’s still all in in Json it’s very hard to read every if you go into a parbi file and change one little thing you could be changing multiple areas of that file all at once and actually I’ve heard some really big complaints from teams that are trying to use this format to to version their files because changing one

14:42 version their files because changing one item in a in a file changes so many different files all at once like let’s say you make a new page or you add a visual or you modify some things right you could be adding multiple artifacts at one time directly into that file and so that potentially becomes a problem problem because when you’re trying to do check-ins or verification of your report via devops or a review process that’s why devop has exist right make a branch do some changes and request that these changes get merged in or let’s

15:13 these changes get merged in or let’s let’s also explain here like what happens when you have two teams working on the same stuff right how do you pick which which individual lines are accepted or not so to me that’s the weakness of the PBP format right now but I think with the timle format coming inside the data model side I think that’s going to I think that will solve a lot of those problems because there’s in the timle format there’s a folder for like tables there’s a folder for measures there’s a folder for the

15:43 measures there’s a folder for the definition of the model so you you have the ability now to pick up individual objects or elements that describe the data model but then you’re now not having to revise one file for everything that is the Bim you now have eight or nine or 10 files and so now you can pick and choose what you want yeah does that make sense for for for some of our audience we’re like yeah Tech talk right like for others yeah this this solution and what’s being built is is squarely in the dev realm I think these

16:13 squarely in the dev realm I think these are the heavier bi teams that require the overhead so I’m I’m happy that there’s continued Innovation and that this is now something that Microsoft is focusing on and Ru is putting this together because there’s immense immense value for some use cases and I think like that that’s where I get hung up and and no I haven’t implemented yet because there’s there’s a for as

16:43 because there’s there’s a for as advanced I think as some bi developers and business intelligence folks who are doing powerbi reporting and building a lot of pipelines and doing all these things they’re they’re not typically also very versed in the the hardcore Dev processes that are in place that take time to to deploy something right like there’s a reason there are processes and Sprints and like when things stop and then whole like

17:14 things stop and then whole like completely different teams come in to do a deployment of code for organizations and this is that process right not it is that process what you’re putting in place here are the checks and balances that take time to review things before they get deployed to an environment and where this makes tons of sense is and and where you might be diving into and using these these features and functionality right now are

17:45 features and functionality right now are where you have customer facing reports where we’ve been clamoring for like a better way to do things there’s just a ramp up period I think because even members on my team aren’t aren’t versed in these areas so so that Tech Deb item I think is something that is certainly on on team’s Radars at this point of we’ve got to ramp up the team we’ve got to figure out like how this is working what are all the pieces do I think it’s like completely there yet no but

18:15 it’s like completely there yet no but like I think it’s definitely in the realm of let’s start to get familiar with what’s what’s going on because at some point there should be an inflection here where maybe there’s a collection of third party tools or whatever to make this extremely easy yes and and simplify that process and that’s where I think this really takes off but at this point it’s just great to see how there’s continued ways in which Microsoft Is providing value is providing the the information

18:46 providing the the information of how we navigate this ecosystem of ensuring we build a quality product which is the report right so whatever the case may that may be I say external Customer because that’s very easy to to say hey like yeah that’s really important but also would it make sense from an internal perspective if you’re the bi team and you’re deploying reports and 90% of your organization is consuming data from them yeah a screw up there is probably

19:16 them yeah a screw up there is probably just as bad if not worse especially around High usage reports right so things like this that add those checks and balances I’m not saying shouldn’t be in business because like of of that situation I just described you want to make sure that you go through some code review and introducing something like this a devops pipeline into your process I I think starts to give you peace of mind that

19:46 starts to give you peace of mind that you’re running code you’re running changes you’re running things through this process where somebody is applying eyeballs on something or you’re testing and verifying that what you’re about to deploy doesn’t absolutely nuke right like a report or introduce something that is potentially misleading like a a Miss in business logic or some additional last minute add down and a Dax equation or something like that yeah and there’s some comments here

20:16 that yeah and there’s some comments here in the chat talking about hey Mike you shared the link today about our conversation and and you’ll notice everything that we’re talking about here is all inside Azure devops yes technically Azure devops is git it’s a repo library in the same way you could do this with GitHub so you could do releases the same way GitHub has a little bit different things on how they Implement things like this thing called GitHub actions but today if you’re going to use Fabric and you’re going to use the GitHub the git integration

20:48 use the GitHub the git integration today there’s no option to pick Azure devops or GitHub it’s only Azure devop so for now everything’s only Azure devops I could imagine ourselves in the future here where this could be in both places but Seth I want to lean on your I want to lean on your your question your comment there talking about this one I I think there’s I think there’s multiple pattern steps here that are I think there are multiple patterns in what we’re trying to do here right so

21:20 in what we’re trying to do here right so there is a let’s think about this in the idea of like certified and non-certified stuff right so I would argue any data that’s coming out of that certified or governed area should probably immediately follow some pipeline or review process to get those reports out so I think when we’re talking about teams that are working on certified data sets Al be this is probably a smaller subset of groups that that are doing things than the broader report of the organization right so there’s for every I don’t know I don’t I don’t

21:50 for every I don’t know I don’t I don’t what the ratio is but like you’re going to spend a lot of effort on those handful of certified data sets and reports because you’re trying to build something something that’s reusable and has the widest audience on that certified stuff that’s where it makes the most sense to spend your time and effort but you also have you you also have probably a lot of other things invested in that right yeah agree yeah there’s things there’s a data engineering team and they’re they’re trying to make sure the quality is in the Upstream data potentially right you have Executives looking at

22:20 know you have Executives looking at things so you got to make sure that they’re happy with the report it looks right yeah agree there but I’m saying to me that’s a good candidate like if you’re if you’re thinking like oh gosh we need to have certified data sets or we’re thinking about certifying things that is that is the portion that that is the piece of work there that would would push me towards this get perspective but Mike have you gone through vs code right now and looked at it data set and like like yes you’ve probably scanned through the files of a

22:50 probably scanned through the files of a pbip in the the project format but have you actually tried in a sense to either go through or edit something or is it really just for the change management and then are you going to another tool that’s the part where I’m this a good question yeah yeah this is a good question so here’s here’s where I think there I think there are two major needs here that are going on inside this example here the first need is we just need backups of stuff right how how can I edit things in the service how can I

23:22 I edit things in the service how can I edit things in desktop and have both worlds have versions of the file and if someone let’s let’s say for example right let’s say for this happens all the time if you are current your current process today would be is I’m going to go have a PBX file I’m going to build it in desktop and I’ll publ it from desktop and I’ll push it to publisher the service everything we’ve been saying up into this point has been the service is a readonly version of the report and we’ve really discouraged people from editing or creating new reports in the service because there were occasions

23:53 service because there were occasions where you could never get that file back out that there there are pieces is where it would only stay in the service now think with me for a moment here our current process is take those PBX files stick them on SharePoint so now we have an extra bit of work around okay let’s go to SharePoint pull the file out do the work publish it again put the file back in SharePoint and now we’re versioning that file last file in wins there is versions of that file but I can’t merge two files together I can’t have Tommy work on page two and Seth work on page three and then merge those

24:23 work on page three and then merge those changes into one file so from a just from a backing up standpoint I think this whole git area makes things a lot easier or adds the capability even for users that are not get enabled if you just imagine I change a file and I check it in I change a file I check it in even just that very simple action of backing things up and having it synchronized to something else I think that’s just in general valuable by itself without you general valuable by itself without not using branching not doing all

24:54 know not using branching not doing all these crazy things I think there’s going to be patterns that are going to need to be developed here that are have not been talked about or communicated yet about git I’m actually doing some substantial work around getting my head around how to best utilize git in lie of the context of fabric workspaces because companies are asking me about how do we build this what does this look like how do you govern this there’s a there’s a very simple use case around just make changes and have it checked in and to the end user all they’re seeing is a button that says hey

25:26 they’re seeing is a button that says hey your files are different than what we have in GitHub type in what what you changed I changed this checkin done I changed I changed this checkin done literally a simple check-in process mean literally a simple check-in process just that by itself I think would be enough to add some value to organizations then there’s a second half of this which is a lot more around testing the quality of the report integrating with and what Ru points out here in the beginning of the article he says this article contains a lot of information there are three open source tools you can use to check things about

25:56 tools you can use to check things about the quality of your of your report so if you really want to go in on it you can say I’m going to integrate tabular editor the best practice rules analyzer the BPA or PBI inspector and these are other tools that you can add to your deployment pipeline that will that will tell you okay for organizations that need a bit more process this bolts on and now you have the ability to assess the report side and the data set side for best practices and I think th that’s something that I think your common

26:27 something that I think your common business user would not need and that’s where my mind goes and maybe like I think you dialed me back when we first started talking about this too right because there is there is a simple approach here there yep purely for that versioning purely for hey Disaster Recovery like you just deployed something revert it it’s going to be a lot faster through this process right or what is it that you need to revert and I I always think about like the extended oh well if I’m going to if I’m going to introduce this process I must mean that

26:57 introduce this process I must mean that I’m going to to introduce like validating and checking what I’m deploying I suppose that isn’t necessarily the case you could just check in deploy right again that’s my point here is there’s two audiences that this is trying to address right there’s the simple just get a backup file and just check it in and then there’s the whole like full boore it grownup process we’re going to have a whole bunch of check like there’s a whole bunch of extra work

27:21 like there’s a whole bunch of extra work that can go on other parts of this if you have reports going to customers yeah for sure those should be certified for sure they should have some standard checks and processes like the fact that we can now bolt on these other tools that are more Automation in nature all this is doing is adding a lot of Rich rule creation or rule automation around the de The devops Experience yeah business users I think would not be comfortable writing even myself like I would consider myself very technical I have not had a lot of

27:51 technical I have not had a lot of success with tabular editor best business practice rules and PBI inspector yet and again I’ll I’ll point out here PBI inspector is brand new this is a brand new tool that’s been developed by Microsoft a Microsoft employee I think his name is that van something can’t remember the name off the top of my head but anyways his Tool’s brand new and and his tool works with building rules about the report here’s here’s a rule engine and you can run those rules against the report s so PBI inspector is absolutely quite impressive and adds a lot of capability

28:22 impressive and adds a lot of capability here but myself I haven’t had a large amount of success implementing these because I haven’t spent the time to really dive in and figure out okay how do I best integrate those tools when do I fail a build when do I fail a deployment because the rules are not met there’s something that’s incorrect In Those Those processes it’s anyways that’s that’s that’s maybe a topic for later or another day but so Mike I think I got something that you’re not gonna like I just sent a link in our

28:52 like I just sent a link in our internal chat and it’s the best practices for the life cycle of just basically dealing with any the tuto endent life cycle management best practices their steps they’re saying with the deployment pipelines is everyone who’s working on a data set or report everyone creates their own workspace yeah in powerbi yes that’s correct which gonna connect I’m not I’m not disagreeing with this so you wouldn’t disagree with this so you’re

29:22 wouldn’t disagree with this so you’re going to have a development workspace testing workspace production and then each person their own workspace okay so so okay so this is this is the point around how does this going to work right so today the way they built the workspaces a workspace is only tied to a single branch called Maine what this is describing to me is saying look if the main branch is the development environment you’re going to be need you’re going to need to take branches of that and do things with it what

29:53 that and do things with it what Microsoft is doing here is they’re basically building an entire they’re B basically recommending if you need to Branch something if you need to do your own work on the data you should create your own workspace named whatever you want it’s like working local yeah it’s like working local you’re building your own separate workspace a local call it a local workspace you’re building everything you need in that local branch so you you still use the the fabric experience you still do all the build that you need and you you grab you make your your separate Branch

30:23 know you make your your separate Branch off of Main once you’re done you essentially would kill the workspace and just delete it it’s just a throwaway piece so everything they’re describing this article the whole workspace that the developer is working on specifically is is throwaway I’m not quite sure how I feel about this it totally makes sense as to why they’re doing it I think I have some mental adjustments to make in my mind because to your point Tommy right I’m not comfortable with having lots of developers just spinning up

30:53 developers just spinning up workspaces but if that’s what they need to build and develop in and and they’re never sharing it with anyone else so why why do you say why why do you say like they’d be spooling up workspace because one yeah because the pattern is you can only have one workspace linking to one branch of your pipeline you can’t easily switch between it that could just mean be on personal workspace you can’t link to personal workspaces you can’t you’re not taking a branch and dropping it into personal work you would create your own personal Dev workspace so separate okay

31:25 personal Dev workspace so separate okay yeah you would create a workspace called Mike Dev whatever and then you would say okay I’m going to go to this Branch or make a new branch on that code and then I would create an entire brand new copy of Dev I could pull all the artifacts down if I needed to where my mind goes with this Tommy is how do you how do you handle data things right how do you how do you handle the data like so everything we’re talking here is the definition of the files and the the objects and the pieces like what happens if this if you’re trying to copy down an

31:55 if this if you’re trying to copy down an entire data Lake how does that how how is how is that supposed to work right you can virtualize the data I could link to the data that’s in in I could have all these shortcuts appear so I’m not quite sure how all the artifacts work yet I think I think there’s some work there from the data side of things on how to make that run smoothly anyways that’s where my mind goes on this one if a data aside I understand why they’re recommending this one it it definitely in my mind it changes a bit of how I would think to build these

32:26 build these environments another pattern that you could use your tummy so that’s this is what Microsoft is recommending here I don’t know if it makes sense to to do that or if it makes sense to have multiple branches where you have a main branch that would represent prod and you have separate branches for test and Dev and in those branches you do the work and then you promote code changes between the different branches until you get to production and then everything gets merged into production so I don’t I don’t know how that’s going to work out yet but it there’s a lot lot to think

32:57 yet but it there’s a lot lot to think through this because this is very new to the the business and I think that’s the word here because like the with the get idea there with the isolated workspace which everyone’s connected to the from the branching like we talked Seth said this a few times you cannot expect people or the majority of like if you say let’s just say I have low expectations you have low expectations you cannot expect people to truly go through and be an expert at get while

33:27 through and be an expert at get while doing powerbi that is such a subset this is great but when you think about the really the opportunities for what teams this these are whether or not You’ even want to implement this right away like what teams are we talking about we’re talking about Enterprise we’re talking to someone highly focused on it would you say we’re working with anyone that’s on the on the spoke or the from the manage self-service side or I because I don’t think so

34:05 the spoke side so you’re saying would would this so if you’re talking about Hub and spoke models or approaches to the to a running organization right the Hub is the central bi team and then you have all the spokes of the Departments that would be handling or managing their own things you’re saying is there any use case for this in those spoke areas is that your question that’s exactly what I yeah so in my two use cases the simple use case of just simply checking things in and out just making changes on files publishing them as normally just checking things in out yes to me that makes sense so the the spokes

34:36 to me that makes sense so the the spokes would use the simple version the simple method to be able to just check things into a get repo on the other side of this I think the central bi team would do the more elaborate process the centralized team would be more technical in nature and they would be focusing more on quality reports Central ized reporting certified things and that’s where you add all these extra PBI inspector tabal editor best business prer that’s where those tools I think apply themselves where where do you see

35:10 apply themselves where where do you see I I can generalize here but but if if if you have a company that is not as far along in their data data Journey right and you’re just introducing powerbi or there’s there’s a little usage right you might have a a centralized team that’s starting to share things with the wider audience is there a difference M Mike in your approach with a company at that level versus way down the road like

35:42 level versus way down the road like high high data culture like everybody knows how to how to generate powerbi reports that are going to create them right there’s a process there’s you’ve got a lot of artifacts you’ve got a lot of great things going for you in regards to a lot of people getting the data they need and developing their own reports right like is there a difference between which of those organizations you would start to start to recommend go down this route or are you saying like right out of the gate you should start training your users how to

36:13 should start training your users how to do this I think the training to do just check in check out going be pretty easy I think that’s I think honestly I think the value of what you’re doing here is good just from that nature and the reason I say that I think that portion is easy to train on it’s literally a red

36:28 is easy to train on it’s literally a red button that appears you C the you also you also then are saying that like they’re not creating their own work spaces and things you have to go configure things for them because they’re not going to understand any of what you’re trying to tell them to do through this process you can’t like I’m not sure I follow your logic there here here here’s here’s my point I’m a business user I’m interested in developing my own reports for myself and my team yep I’m brand new to powerbi okay I can

37:00 yep I’m brand new to powerbi okay I can click publish that sends my report to a place I can share it I’m done yep no problem why do I have to go through all these other steps that you’re telling me to do why you why do you keep thinking such a hard weight I think you’re overestimating how hard this is it’s not hard you’re overestimating the whole idea so you okay let’s let’s extend your extend extend your scenario how many times does that business unit to need to figure out that someone overwrote the report and they blow away all their changes before they start oh that doesn’t work well how many times does someone go in and edit the report in

37:30 someone go in and edit the report in it’s not the they like the likelihood of somebody else touching their report they don’t they don’t work on a team this isn’t a bi team this is a person there’s there’s going to be a team of people so okay let me give my observations here then my observations is there’s very few workspaces that just have a team of people in it or a person who’s using a workspace to share things to their group most business units do not understand this almost every time I walk into an organization they just add had a crap ton of admins and members into a workspace and they’re all mucking with

38:01 workspace and they’re all mucking with stuff all the time so one that’s so one first thing is if you’re giving business Union work spaces like this and you’re just saying here you go and they don’t have any process or plan around go who’s what in that workspace one that’s a miss so already that’s a Miss from the central bi team they didn’t do their job on educating the team as to what is an app how to distribute content and how to get the right users in that there should be one admin maybe one member as a backup and then there should be a whole bunch of people as contributors and

38:32 bunch of people as contributors and contributors are not there to look at the reports they’re only there to make stuff they’re there to build things so to me you’re already you’re already missing the first point there is because you’re you’ve already put question I’m not missing anything well I’m just saying asking a question so the other thing here I’m looking at too is if you’re going to be doing this many companies have and I think someone in the chat was Darren was talking about this as well their company doesn’t let their teams make workspaces so when

39:02 let their teams make workspaces so when you’re so there is this idea and I have a work with a lot of organizations that they don’t let people make workspaces they’re going to give you the workspaces and you have a form to fill out to request your workspace so in doing that there should be and I think in processes that I’ve built for companies there’s a training exercise here’s what the workspace is here’s the here’s the rules of the road oh and by the way since I’m Central bi and I’m setting up your workspace I’m going to turn on this thing called f Fabric or not I’m going to turn on git or not if it’s in premium

39:32 to turn on git or not if it’s in premium and then you you now train that team and say look we’re going to turn this on here’s how you use this and then if they choose not to use that no problem not not an issue but I will set up a git repo to that workspace I would set it up from the central it space because that allows me allows me to by default track and store changes on things so that that way they’re not that way there are when this occurs because it will occur someone overwrites

40:03 because it will occur someone overwrites some data something goes bad or someone edits a report in the service and now someone in desktop is trying to get that information back down there’s going to be some disconnects there that are going to cause some challenges and problems and so to me I will work on that educational part just to give them the bare minimum to integrate this because I think the upside of this as they become more comfortable let’s imagine 6 months a year the road down from from now there may be actually some really solid use cases around that

40:33 some really solid use cases around that business unit just having that built in I I agree I think I think the value add here too though is and something to think about even in in those different Arenas is you’re you would already be training those business uniters on a whole set of processes prior to that which is like okay the only way you get Version Control is if you if you save your file in SharePoint and then when you’re in SharePoint totally you’re do steps are you already doing and introducing to get that level of like Version Control and if we can put that

41:03 Version Control and if we can put that up near the place where you’re doing work you just don’t do anything you don’t deploy anything unless it goes through this process and that’s the last check that we’re going to do before we send it it make that makes a ton of sense and I and I see I see with this feature so I could see the writing on the wall here maybe and maybe you might disagree you might agree you might disagree I think with this git integration you could essentially kill the whole SharePoint process I think you get rid of it I think I think this deletes the idea of having individual files saved inside SharePoint this this

41:35 files saved inside SharePoint this this is no longer needed the only the only side argument I would have with that is I I think the centralization of PBX files is an important thing like every everybody on teams or everybody in a location should have access to the things that are out there totally agree but that’s the whole so I agree with you Seth 100% but that’s also so things yes but you rely upon the user to correctly store the right files in an organized way to

42:06 the right files in an organized way to do that 100% you are yeah but what I’m saying here is I I that statement I love your statement it’s so good the git integration makes that even easier because now you actually have control so imagine you have hundreds of workspaces hundreds of workspaces could be centralized into one git repo or you could have inside devops you can have these things called projects and you can build projects out by individual teams and so now now you really do have a central trackable location for all those

42:37 central trackable location for all those files no matter what and whether or not the user follows it or not as long as gets turned in but as you all the files are stored and and there but if you you extend that too then you really don’t care where people are building your reports because potentially like we could be building reports in the service here pretty soon what if we have so at that point in time at that point in time when you run into the company who which of you ran into it where everything they were building was in the service get it back in get it literally get it like get it but but that’s but my

43:10 get it like get it but but that’s but my point is like I think to me to that point like right now we are really talking about centralized reporting now we’re really talking about okay request request a workspace we set up the G for you it atically happens to the business unit they don’t have to do anything all they do is just create reports publish things no problem and then the the the amount of amount of training comparatively right SharePoint store files versus check in checkout in the program that’s already there with a couple button clicks I would argue it’s

43:40 couple button clicks I would argue it’s actually easier to use the get integration than to start working with people to like where to upload and download files from yeah so that’s kind download files from yeah so that’s my main argument get on the of my main argument get on the train so that’s why I want to perspect give perspective to this like there is a simple and there’s a complicated version of this and so now the game is okay cool we understand the simple version simple version is built into into powerbi. com love it what about the what about this next stage of the game like what what

44:10 next stage of the game like what what happens when we’re talking about PBI inspector and tab editor and best business practice an analyzer those tools require so much it support who’s going to come out with a solution that makes that part super easy and clickable just like in desktop to me that’s going to be amazing because that’s going to what that’s going to do it’s going to take a very technical topic and bring that closer to where those simple business users are simply using checkin checkout so I think there’s a path here

44:40 checkout so I think there’s a path here where techn get better yeah well let’s hope right right but I but I think I think that’s my point I I think the the thing I love about this right is this this opens the door to a lot more things we can do and if we know any if if we know this but everybody else may not like Microsoft is an enabler right like they don’t build every single feature or all of the UI whatever like

45:10 feature or all of the UI whatever like there’s bu ecosystem yes of businesses and it’s large part of that as Consulting but of tools and third party thing like there’s an ecosystem on all their products where there’s this wide space of these solution that simplify a user experience or provide insights that you could go custom build yourself but they this this other

45:35 but they this this other company did and I think like to your point Mike this does open the doors to a a a huge feature capability yeah it’s just it’s just how do you engage more users into it I do I do agree with you from the standpoint that if you’re going to teach one process and there’s a simple path into this then that would be something that you you would probably you’d want to introduce early and often so that you don’t find yourself in a catastropic phase where it’s like somebody deleted

46:05 phase where it’s like somebody deleted my something my report where where’s that last version where’s the where’s the one that you deployed I don’t know that was Sue she left three months ago right what SharePoint was it which version of the file was it like I don’t know like and and now you need things like okay well what what’s in the service and did someone edit it are people now editing Sue’s report in the service where did that come from like so and I think it’s it’s the the the win on the business side is going to be is it still easy to do is it still simple to

46:35 still easy to do is it still simple to get my things out the door because I have less interest in your technical jargon right I don’t care about your versions of the things of the whatever until I do yes and in it the vast majority people have seen things deleted and just go oh my God that’s like another week’s worth of work right or those catastrophic things happening are on a wider scale because they impact more of the organization so that’s why those processes are in place and why the organizations like I am 100% okay with

47:06 organizations like I am 100% okay with that overhead because we’re we’re making sure we we’re double redundancy right like redundant redundant redundant so that we’re always safe and running well whereas the opposite side of that is the the conversation to some degree we had on Tuesday just the differences between the priorities of business versus it but as as we’re pushing more of these Technologies and things frontf facing into the business I I do agree with you I think the win is

47:36 I do agree with you I think the win is where it’s it’s simplified enough where yes it’s part of a process but it’s not something I have to go spend a ton of time to learn or understand I’m just following it the real benefit the real benefit though is of these things for the the Bic or the teams that are managing a lot of this is as at like one of the things we talk about a lot is reports that are crashing the system like just poor practices that are being deployed and costing the money or costing the

48:07 and costing the money or costing the business a lot of money if you start to put Frameworks around these things it’s very easy for the technical teams to come in and integrate or come in and see what the problem is or look at like all of as these things enhance are the rules being followed are you like whatever other cases there are because we now have insights into this file and the all of the artifacts within it you can triage and figure things out much faster and I also think it would

48:37 much faster and I also think it would also it eases that path of like how do you go from a personal report or a team report to a much wider audience cross departmental large scale within the organization because that should follow a curve or a process into a team that manages that yes exactly and this is where this makes a ton of sense where it’s like okay great now I have visibility into what you’re doing we’re going to pull this up and in and we’re going to run it through a a process or a series of tools that allow us to

49:08 or a series of tools that allow us to understand like what you’re doing because that that has been the conversation for forever in that like how do I go in and enhance or tweak or refine or performance tune or do all these things on a report because you’re going to need to do it on every every single thing that comes up through the business just remember you need you need to build you need to build all the things for just $12 right from from our example yesterday build all the things add all the G do all the check-ins but just do it for $12 a month like it’s it’s like this idea of

49:38 like it’s it’s like this idea of like I want everything but it’s got to be be cheap and and that’s the thing though it’s like as long as it’s a one two three click right that’s fine I’ll follow I’ll follow your thing exactly right and and and and then and that’s really really my main point my main point is right there’s like there’s an easy version of this one and there’s like up what is it what is in Duke Nukem there was like there was like two versions of like in Duke Nukem there was like baby mode and had like the the guy had like a little Nookie in his mouth or a pacifier then it had like Ultra hard and he’s like War mode like I

50:10 Ultra hard and he’s like War mode like I feel like this is one of these situations where it’s like okay there is an easy mode to this and then there’s like epic mode like and like let’s do checking of reports let’s do automated everything like extra hard mode and I have found when I do automations and of things like you get to that point where there is ultra hard mode on things like this and you just got to spend time learning it to get comfortable with it and and not it’s not a normal business user mentality right it it it some

50:41 user mentality right it it it some business users will like it and want to do it others will not and and a majority I think will not so this is a but this is a good story though that’s an that’s an easy that’s an easy win for a Coe or like you’re setting if you’re setting a process and the business is like nope we’re not doing it like no you are and and as long as you have a like sponsorship like that should be a no-brainer like correct why would you be forcing these people to do it because with four clicks we ensure that they’re not going to destroy their lives oh okay well that makes great

51:11 lives oh okay well that makes great sense do it exactly yeah and and also it provides visibility to the entire or so now you have all the richness of whatever all the code that lives inside your Azure devops so if you ever need to go get access to multiple teams reports across like so let’s imagine there’s a problem in some team and you’re trying to support them right it makes it much easier from a support standpoint to show up and say okay what workspace are you okay yeah I know okay let me see okay what G are you in okay great you can go

51:41 what G are you in okay great you can go right to the git and you can pull down those reports individually and say okay I can see what your problem is you can go go to their problem specifically you don’t need access to SharePoint you don’t need access to other stuff everything’s in you can now have oversight of all the reports across all the the things so many so many doors to open so many doors yeah so many doors all right with that I think we’ve definitely exhausted this topic around get and get replacement pieces I think

52:12 get and get replacement pieces I think there’s going to be more coming for this I also think there’s more challenges coming with this depending if you’re Pro or premium user because I think you can use git in both situations but the pattern in which you use it would be slightly different so there’s going to be different pattern I think as the licensing applies here regardless this just means we have consulting jobs for the next foreseeable future I think we we’re good like thank you Microsoft for giving me a career that I can continue to invest in long term you can continue to invest in long term I appreciate that but with that know I appreciate that but with that let’s wrap thank you all for your

52:42 let’s wrap thank you all for your time the comments have been awesome so thank you everyone for commenting and chatting things here some of them are really funny so I really appreciate the funny comments I enjoy a good laugh in the morning it helps with my therapy session that is the podcast so if you like this content if you felt like this was a a good topic to start getting your head around this new thing I recommend share it with somebody else impress your knowledge to your friends by talking to them about git and how well git from listening to the podcast like I think we should introduce

53:13 podcast like I think we should introduce get maybe maybe I wouldn’t go that far but a new buzz word we do appreciate your listenership we do appreciate everyone sharing this as much as possible so we we appreciate if you like this conversation if John KY please send him this episode I’m pretty sure John kky would love to jump in on this and have some comments here later on he’s he’s like Mr devops for powerbi he’s like all about it so get him involved here as well if him send him a link to it and bombard him with links it’ be funny with

53:43 with links it’ be funny with that being said we appreciate it please share Tommy where else can you find the podcast you can find us on Apple Spotify wherever get your podcast make sure to subscribe and leave a rating helps us out a ton do you have a question and I idea or a topic that you want us to talk about in a future episode head over to powerb tips podcast leave your name and a great question finally join the conversation live every Tuesday and Thursday a. m. Central and join the conversation all powerbi tips social media channels awesome thank you all so much appreciate your time we’ll see you

54:13 much appreciate your time we’ll see you next [Music] n

Thank You

Thanks for listening to the Explicit Measures Podcast. If you enjoy the show, please leave us a rating and review on your favorite podcast platform.

Learn more about our work at Explicit Measures and PowerBI.tips.