Automate It. Episode 13 - Power Automate Desktop Monitoring Dashboard

Automate It. Episode 13 - Power Automate Desktop Monitoring Dashboard

Show Video

What's up everybody? Welcome back. This is  Automate It episode number oh i gotta go two hands   thirteen. I am Jon Levesque, i'm your host and as  always i am joined by the awesome crew of Kent,   Pranav and Apostolis. Welcome guys how you doing  great john how about you doing good doing good. Kent how are you up in cold Canada uh not too  bad not too bad i see you got a fancy new studio   behind you that looks pretty cool yeah yeah i  got an upgrade in digs we're no longer in the   office i'm now in the studio and pranav how you  doing buddy i'm doing well in the rainy seattle   show i know i wish i wish we would get some snow  or some sun or something else i'm tired of gray   and wet i can sell you some snow that's not a  problem i mean a cooler of snow is not enough   Kent i need like i needed to come from the sky  you know i appreciate the thought though i wish   the borders were open i'd come hang out up there  and and also you know what real quick let's say a   big happy new year this is our first episode back  after the new year happy new year to everybody   glad to have you back with us yes okay so on to  the business at hand today Apostolis is going   to show us some great stuff for so many of you  the out of the box analytics you get in flow and   power automate are just not enough we hear all the  times how do i get deeper how do i find out more   information and so apostles today  is going to actually walk us through   how we can do just that how we can dig into  the common data service find the real set of   statistics where all that information lives and  decide what we want to display for ourselves so   apostles i'm going to go ahead and hand things  over to you my friend and the the floor is yours   yeah thanks john so the topic of today is really  how as john said how we can go deeper on the flow   analytics uh now that we have uh you know the to  distinguish also between cloud flows and desktop   flows there are of course different reasons why  you would like to dig deeper on specifics uh of   uh past flow execution or maybe even on real time  so in order to uh enable that um we leveraged   uh something which is called microsoft databurst  this is the underlying foundation of course   of the power platform but also uh driving the  whole dynamics 365 business solution stack   so we will be leveraging uh the microsoft  dataverse api system together with power bi   and the power ultimate service to get a little  bit deeper when it comes to the flow analytics   so for that let me actually jump into the  deck so the agenda will be we first of all   the key drivers why why we did that dashboard  how it looks like how it has been built some   really important findings around that and then  looking also into options of real-time monitoring   and of course q a so before we start what i  just said the underlying foundation of that is   of course uh leveraging uh the power platform the  breadth of the power platform um to to you know do   this low code no code kind of extension and even  reporting to derive those insights we are looking   for so in our scenario from the left-hand side we  will be using power bi the business analytics tool   power automate of course for the flow execution  and the Microsoft dataverse underlying foundation   which is uh much more than only a data store right  it's really uh it is a hyperscale database as a   service if you would like so if you're looking  here what this is comprised of i mean this is as   you can see here on the right hand side Microsoft  database first of all it runs on azure and it's   much more than only you know azure sql database  so it is really comprised of many many components   here making it super scalable highly secure and of  course hyper uh hyperdense and stuff right so and   if you're looking at the internals from a feature  perspective of dataverse as i said it comes with a   lot of built-in services like security logic data  storage integration and then if you're looking   into the you know the foundational topics like  authentication authorization but also auditing   so you know who who's accessing those components  when and uh you know have the full traceability of   how many executions of a specific activities  on a table has been done and so on then on the   right hand side you see many connectivity  options reporting you know connectivity to   azure data lake and so on so forth so and the api  this is what we have leveraged in the scenario to   derive to connect to those endpoints which are  you know whenever you execute a power automate   desktop flow uh the telemetry the execution lock  is then of course transferred to our cloud service   and from there we can access uh that data  through the data verse api connectivity   and this is of course enabling many many scenarios  across of course the power platform and even   beyond that when it comes to integration and to  surrounding of of the data uh sitting in database   okay so why this dashboard first of all i was  bored during the lockdown uh i mean yeah you   know it's it's it's it's easy how it's really  amazing how easy you can get started with those   uh components once you you know a little bit  the workings of the power platform it's so   easy to extend that knowledge to different  uh tool sets like powering bi in this case   or even smaller power ups which then surface  information uh or even allow you to to uh to you   know enter information which will then be part of  the overall flow so it's easy to up skill on that   and uh you will see this uh in in the real-time  dashboard at the end as well so it's focusing on   uh desktop flow uh insights what i said before not  only uh when an automation has been executed but   how long it took what actions have been called  uh what actions in in that automation have failed   uh which ones uh you know took longer on the  previous week for instance or previous execution   some statistical process control you can do on  these these components and these statistics so   that was really uh to get deeper insights on on  what's going on from an execution perspective   um showcasing of course the api endpoint  consumption what i just said so data versus a gift   that keeps giving all these built-in hyperscale  uh you know architecture security by design   record level security so you know whenever you are  you know accessing the end point it's respecting   also the record level security for your user uh  which has been applied audit management and so on   then it allows for reactive governance and control  like we have with the coe starter kit where   you can really uh control execution on action  level if you would like to we see this later in   in the chart and then you know proactive uh reach  out instead of blocking the whole service just   proactively then monitoring and reaching out to  folks and ask okay what were the use cases why   they have used certain actions just to understand  a little bit more about the demand is within the   organization and also help drive of course the  citizen development ship identification of api   based operations and also pro code actions we  see that in a second what that means what i said   before statistical reporting so you can really  compare flow runs with previous florence and then   you know judge if this is considered an outlier  or if there's a pattern you can see okay   where you need maybe to look at the  hardware or the vms and and so forth   it is easy to deploy this was also  you know one of the design principles   many organizations they they don't have uh you  know the full breadth of the power platform   components enabled today or deployed things like  the coe starter kit which is really a massive   governance framework so if you just get started  with the you know power automated desktop and the   power domain service in general and you don't want  to have uh the cre starter kit deployed in your   organization but you would like to get also in  the same time of course those insights from from   the power automate service this is where you can  really get going with this starter template and   the last point what i just mentioned is  really for organizations that haven't   uh you know enabled the breadth of the platform  yet so all you need is of course uh access to   the database uh backend sufficient privileges  is very important because otherwise you won't   see uh things uh power bi desktop and off  you're going okay any questions uh john team   no it's all very clear to me so i guess one  question just around uh when you mentioned   the right level of access uh so what is required  in order to to be able to retrieve this data so   first of all you have to be uh you know flows  are assigned to specific owners so whenever you   you author a flow you are the owner that of that  flow and then you can share that flow of course as   co-owners with others so this is very important  that you have to have uh dataverse environment   access first of all so if you don't have access  to that environment you would not see anything of   course and then it goes really further down that  you have object access you have rbox on the cds or   in the database and tables that you can say okay  on a flow session or in this and that workflow   entity i would like to restrict that access to  that so that's the idea behind yeah and whatever   you are accessing you cannot even restrict it  so let's say a flow has not been shared with you   and also the execution has not been shared with  you that normal regular user which is not really   having any administrative control won't see that  right so you don't get more access by accessing   those power bi report through the dataverse api  system because that's the essence what i said   before it is secure by design so i guess just  to follow up so if i was a regular maker that's   building say api flows or cloudflows and desktop  flows i would be able to see my information but if   i say was an administrator i would be able to see  um more information because i've been granted that   administrator role is that a is that accurate then  that's correct and you would see i will demo this   at the end as well so what a normal user would  see and what the administrator sees and that's   i think one of the essential parts why i said you  know dataverse is the central piece of our whole   power platform and it is so important but that  always front and center because all these you know   uh scalability and security configurations and the  auditability of that system this is just uh god   given for us right as soon as you have a license  for a power automate desktop or for power domain   you can leverage all these technologies which are  out of the box you don't have to worry about them   right it's always whether you access the this  data through the api through the ui or other   means it's always respecting the api and the  security layer behind it so that's that's very   powerful of course yeah i like that it's god-given  rights as soon as yeah that's awesome that's great   all right so let's have a look uh at  the at the demo let's do it demo time   so first of all i'm opening here power sorry power  bi and which is leveraging what i said before the   Microsoft dataverse api to get all this  information here so we're starting here   and these are this it's on purpose renault you  know it's a starter template so this is meant   to be customizable like the rest of the series  starter kit you can really customize this to your   needs and here's the data model of that and here  you can see a graphical representation of that   uh you can really uh change uh you know add new  relationships to that you can add new measures new   column calculated members and so on so once you  have started that as a template you are asked to   provide two things one the dataverse environment  url and the flow session history url in poweradmin   portal then you have to authenticate with a user  that has proper privileges and then you are print   it's loading the data and it's presenting  it in in this starter dashboard here okay   so from here what we see is some filtering section  here on top where you can flow status you know   failed or succeeded run mode attended unattended  or executed from the power automate desktop   console then some basic statistics here on top  how many makers do i have in my environment   how many desktop flows how many legacy ui flows  well which is using you know still the windows   recorder this we come in a second to that so how  many overall flow runs i had uh in that time frame   here attended unattended runs console runs how  many bots uh how many canceled flows failed and   so on and this one is uh something when you have a  super large power automated desktop runs which are   running and many loops or a large loop within  those uh it is producing a lot of course log   contacts and this to identify those large log flow  runs this is where you can see them here in this   overview then you have a couple of uh lists here  top three makers top ten bot hosts those are the   hosts where the power automate desktop have been  run on through the on-premise data gateway uh   you know total hours processed and so on  uh then the top actions in power automate   desktop which have been used along the line you  can see of course the right to excel has been used   23 25 000 times what was the average uh processing  time on it so just a couple of uh you know basic   statistics uh you can get all basic counts you  can get out of whatever the flow portal is really   producing here from an action perspective  so if i click here on one of these flows   and i look here at the processing history so  all these actions which are registered here they   are compiled in this power bi dashboard rights  so from here as i said you see action subflow   start duration and status but when we are looking  later on in the action details you see of course   many more things let's go to the desktop flow  overview so this is also another important view   which gives you on the desktop uh flow name level  right here it gives you really a quick start how   many times it has been run failed importantly how  many action counts uh have been included in those   flow runs uh is it using proscript we come to  this in a second how many uh you know action   calls have been done with pro script or for  proscript and what is the total run time on   minutes for that specific automation then we  have here an indicator whether this is a legacy   a ui flow or not uh just to give you an idea  okay which of those might be potential targets to   migrate to the new power ultimate desktop now what  is very important here to note is uh this this   feature here and that's this is where we start to  really dig deeper on on those flow runs as you can   see here as well in the overview statistics  so i have here an indicator which says okay   how many flows are using uh those pro pro  scripting components in my flow run history   and this will give you a good indication you  know how many of these have been uh including   things like run a vbscript run powershell run  python scripts or open a command session so on so   these are usually indicators where developers rpa  developers if they're citizen developers or not   when they leverage those you know more advanced  components in the power automated desktop then   you might want to dig a little bit deeper without  opening each and every uh flow individually you   would like to have really 30 000 feet you okay  how many of my desktop flows are leveraging   technologies which might reach out to other web  servers do some you know uh powershell scripting   which is of course very uh very sophisticated  things you can do and potentially also uh   dangerous things you can do with those components  so to get an overview of those this is where these   kpi here comes from that was the idea behind it  right and we can drill down into these in a second   so overview starts desktop flows what i said  and now the desktop flow monitor the desktop   flow monitor is nothing different than what we  see in the power automator portal and also the   layout is aligned to that so if you look here back  uh the status is what you see here and start and   stop and duration and so on so here conveniently  we are showing also if it has been a failure   what the error message was right and from here  as well you have two links you can go to the   actual uh power automate desktop flow so if i  click on that so it's opening the power automate   portal and directly goes to that position of the  flow xor to the to the flow run history of that   particular record and here i can again review  whatever has happened and the error message of   uh that thing so fail to obtain output for custom  object here for instance right and here i can see   the screenshot as well and what has happened  exactly so if i minimize this or i can go even   if this automation has been triggered from a  cloud flow i can also go here to the para flow   and then i have a look okay what happened here in  that if there was a failure or not or white has   been cancelled and so on so it's opening again on  the other screen that's handy to give a view that   looks aligned to the portal you know something  familiar so it's like oh this makes sense to   me i've seen this before exactly especially  uh john in case of cloud flows you know that   we have this option you can see each and  every component in the cloud flow so we can   really debug that specific uh action uh which was  causing maybe the error or even you would like to   look at the input and outputs of those actions  individual individually just to make sure okay   you understand what what happened in in those  environments okay totally so from here on what's   important as i said uh in the cloud flow history  you will see if you go back to the history records   you will see here a note which is saying on the  cloud flows of course not on the desktop flows   so you will see a note here on top of the run  history which says 28 days run history so anything   you know beyond that is cut off because of  gdpr and privacy requirements and so on and   to indicate you know before you before you really  click on one of those icons it will give you an   indicator graphical integra indicator with a flag  okay there is uh that process execution or that   flow execution is older let's say than uh 28 days  okay so here then you see then the indicator of   those flows okay so let's go up here and look  at the next component from here on and that's   also important right if i would like to review  of course my flows and i would like to look at   more details on specific errors what is nice here  what you can also not do from a portal side is   that you go for instance here you say okay on  letter specific automation which succeeded in   this case i would like to right click if you  would like to drill through that record this is   really the execution of all flow history right you  can now go into details like to the run details   so which actions have been included as part  of that execution you see here 24 actions   the average duration of each action was 1.85  seconds and what was the max duration uh of   a specific execution here okay so in here you  can see even again the flow itself so it will   open again the portal or you can also look at to  okay what information drilled down further what   information uh has been used or what actions have  been used uh in in those flows if for instance a   script has been used then you would see you see  that here or if web service calls have been done   through that then you will see that in  this experience so let's have a look at   uh actually scripts i think those will be  used here okay oh quite a lot of scripts   as you can see and this is now i think one  of the key reasons this has been actually   built this gives coe folks a center of excellent  folks really a view into uh the situation across   environments or in this particular case of course  for that environment but for all uh power to make   desktop flows and their respective actions so if i  go back here as i said here you will see a subflow   the action name which always is reflecting  here in power automate desktop so those are   the subflows and those are then the  specific actions what we see here   is exactly the detail level and then the duration  in the status but going one step further if you   i would like to have a look here for instance  if we go back to our power to my desktop flow   and i look at one of the uh really pro code  actions like you know calling web service   so if i double click on that invoke web service  action uh here i see okay i have a streaming url   and here i have the request body and  the request body is actually at this   json object so i can i would have to go into each  individual flows and then review that stuff and do   my auditing and so on but i can uh by accessing  the dataverse backend as we said before i can   conveniently conveniently do this directly from  this dashboard experience if i go here to scripts   and i drill through here we'll see okay we had  a write to command session action called which   was calling a host name then there was a python  script executed from dos command we had a run   application action which was then calling python  and then obviously some parameters uh powershell   scripts here vbscript some javascript or  even python scripts verb as part of this   specific flow here right this gives a really  a huge visibility over the whole uh command in   action history if you would like to look at that  on a general you know on a higher level so for   all flows then there's also the script action  monitor which allows you to filter that by a   specific desktop row or even an action name if  you're looking for only bb script for instance   or run vbscript you can filter that here of course  and then it will show here below all actions which   involve vbscript in here for instance this is a  an action uh which is calling some sap automation   and then you can review okay are they doing  something specific something complex here or is   it using what transaction code is it using and  just you know then maybe to proactive approach   again those folks say okay we might have some  standard interfaces for that or maybe an api   so just to to drive that discussions and have that  visibility what do you think folks i think that's   a lot of scripting that's amazing though that you  can dig into those individual steps and and show   every piece of information that's passing  through that desktop flow it's pretty awesome   indeed i mean that's you know the kind of  visibility uh the organizations or you know   larger enterprises at least look for when they try  to enable the service for the whole organization   and having a reactive monitoring kind in place  which gives them really uh a lot of flexibility   when it comes to auditing and reviewing of  those processes okay so this is from an auditing   perspective so also seeing and you can really  define i've just filtered here so if you open that   thing here i've just selected which of those  actions i consider to be pro code if you want to   debug two through you know with other ones if you  want to see okay how many times a write the exit   command have been called you just i can filter  that and then they will show up in the respective   uh lines here so this is as i said completely  flexible and up to the consumer to adjust   to their needs so the next level will be okay now  i know what actions are you are being used how   about looking at uh you know process variability  and also execution variability so that you know   you see early signs if things go wrong or some  of the processes run really longer than expected   or they are failing consistently or  consistently over a period of time   so for that purpose you can hear filter then  for instance if you have a particular one   let's say desktop flow process vendor invoices  and then you can limit also the amount of uh   so the time frame you would like to look at let's  say then you have a week's overview of the data   you would like to have a look at and this is what  the if you know the the chart would look like   with the information you have filtered now  here from here on you can then say again   right click and look at that particular day  the executable execution of that particular day   by drilling through to the details and here you  come again to that view which is filtered now for   that specific uh process cluster on that day all  the execution and here you see immediately okay   there was something wrong here as well as well i  would like to look at that look at the scripts and   then see okay there was some some run application  happening only some system in for who am i and so   on you can read it and you review and look at the  error so from here on as i said before you have   always a link back to the power automate portal to  look at the specifics for that specific execution   and here you see if i scroll down i can click  on that one there was a run subflow and this has   failed actually the real action which has failed  is further below here and this is the problem so   this gives you really you know uh a link always  between your custom reporting you can always play   back to what the reality is of course in real time  on those power automate desktop and power portal   statistics okay and from here again if you would  like of course you can then drill through again   to web services and so on back in the performance  chart as i said this is the historical view you   could get and then you see of course here always  the duration and the this this bar here next to   it always reflects the previous days uh or the  previous periods uh execution count all right and   this is then the distribution and the change over  time you see there as an indicator of course this   one is maybe not a stable process it might have  you know different inputs and outputs so that's   why the processes run more often or longer but if  you have really stable uh process with a stable   variability of an input and output this will give  you really good insight on on what's going on   then the next level if you are sufficient in you  know analyzing those those points then you can   let's filter here to succeed only you can also  look at a more statistical approach on that data   right where you see here let's start from the left  hand side you have here upper and lower control   limits and then you know a central point and this  is then controlled how many standard deviations   your process is allowed to deviate from from the  mean right so if i want to control my process with   uh three standard deviations or sigma's i can open  that a little bit okay i can say this is still now   under control my process or if i want to uh drive  a little bit more narrow uh execution and you know   tighter execution i can even go to one or two uh  sigma level below and this would then show okay   i have an outlier on these on these dates these  three points here and then you can look and dig   again deeper on on this side okay and here's just  the violin chart which shows you the distribution   on the different hosts you are having where where  those particular flaws are run on make sense   yeah the level of intelligence here is intense my  goodness i think each time you click down i'm like   how can it get deeper my gosh yeah i mean you  know and that's really not rocket science uh   john right so this is normal and that's  the power of power bi and actually of the   dax language so uh this dashboard has been  pre-loaded so i've done a couple of these   measures as they are called which are really  giving you the opportunity to look at data   from yesterday for instance previous quarter month  over month changes which we will see in a second   so those have been defined and you can  see here the code you can change this   but it's also a good template to learn maybe those  technologies and how we can use that you know   without any coding or statistical uh backgrounds  because powerbi and the dax language give you a   lot of uh things out of the block where you  can create for instance new quick measures   which have a lot of this intelligence building  so you don't have to think uh or to learn the   language really from scratch right where i see  okay average per category time intelligence or   uh you know month over month changes quote up  or quarter changes and so on so that's all built   in again in power bi uh so there's tons of videos  of course available in documentation on how to do   those but that chart is meant to show you okay how  you could address those specific reporting needs   uh if you wish to okay so this is then the view  you see on the on the portal when you have a flow   execution right so header flow execution these  are the actions and then durations and the stables   so if we look now at this action performance day  over day or month over month this will give you   a different intelligence now again on action level  but it will tell you also for that specific time   frame uh what is here for instance the previous  execution length for instance duration for   vbscript for instance and what was the execution  of the previous day and then it calculates the day   over day changes right and you can look at this  on action level and of course time period level or   by host by action and so on so forth and then the  same goes for month over month changes where you   see okay for specific actions let's say you always  having you know an update update script which   does some uh month end uh transfer of data and you  know always okay it's the same amount for instance   stable amount of employees and a stable amount of  data which should be transferred you can see okay   performance of that specific action which is then  calling maybe a web service on on the other end   and then you can look at that from uh from  from a really uh month over month shade   day over uh day change and and so on okay  and then the other three those were just   the drill through uh reports you have seen uh  before so this is this is all good and nice so   imagine uh you have a power automate portal for  our analytics where you can see not only of course   the actions as you might have noticed we  have here a filter which is also hostname   which is actually the on-premise data gateway  or a cluster and the same information we have   here of course also the portal right if you go  here to our data and then on the gateway side   we see a list of on-premise data gateways or  actually bought bot hosts and for targets so   single on-premise data gateways or here we have  two clusters one is a windows 10 cluster and this   is a windows server cluster right so if i click on  that then i see i have two participating clusters   in that gateway and here i see again per club per  a machine or per cluster the previous execution   of my pad flows okay and here i did the three  different sessions so that's that's good and   that's nice uh the problem is let me start one  of that of our cloud flows which is triggering   so that cloud flow you will see in a second is  triggering in parallel three uh power augmented   desktop flows which are executed on a specific  cluster this server cluster i've just shown you   right it's executed on that cluster and because  i have set here that flag run on all gateways in   the cluster it makes a choice depending on load  balancing where to run uh the the bots okay so   these are the three uh uh bots i'm having here or  the same bot sorry but about the three executions   of that so i have three accounts here which is  then accessing those machines uh accordingly so   when i start now that flow and this has been newly  added in December so they're really monitoring   a cloud execution on on gateway level  so if we go now here let this start   okay so this has started as you can see here  so if you go now to the gateway overview   i go here to monitor and then desktop flow cues  what i will be seeing here is then first of all   again my gateway cluster gateway a gateway cluster  how many gateways are part of that cluster and how   many if i go to the live updates here how many  of uh those clusters or gateways are currently   executing flows and here we have two desktop flows  which are cued and one which is currently running   so if i click on that you will see really again  the live update uh depending on priority uh if we   go back here these three one has called with high  priority high priority normal priority and this   is of course then reflected here in this queue  as well the problem now with that is that that's   good and nice but you don't see uh okay which  machine has been executed on you don't see what   is the anticipated duration of that specific flow  that happens here and what is the current context   let's say you have you know a month and closing a  power automate desktop which has really long long   running reconciliation process and you have a loop  with maybe a thousand records and so on but it's   really running maybe two three hours it would be  very convenient to have a live view or at least uh   uh you know snapshots of the current execution  state so i'm in the 15th loop for instance   and i'm currently working with that category  of data so in order to enable this again power   automate sorry power bi has been uh used together  with power automated desktop to define a so-called   real-time streaming data set so what i've done  here in power bi in the service i have created   a real-time streaming data set which has these you  can freely define them you can add and create and   you know just as your data schema requires i've  just defined a couple of properties here on the   real-time streaming data set which allows you  to collect information and save that either   so if this plug is on then the data is stored on  the power bi service if this historic flag is not   on then the data is only cached for an hour and  then it's gone so this is of course uh you know   for purposes like iot monitoring and so on where  you don't have to store maybe eventually data   but you would like to detect real-time anomalies  and so on so that's what i've done very simple i   defined a couple of properties here and then i uh  this is generating automatically for you an api   you can post the data against right so here is  the body of the data if you're using a c url   or powershell you can do that accordingly okay  so then what i've done on powerautomate on that   specific so you would do this live monitoring  of course only for specific boards you don't   want to have that for bots which are running five  seconds or 10 seconds or really not critical also   from a sequence perspective but if you would  like to do that then you could send uh those   you know monitoring uh events or those status  events to the power bi service to the streaming   data set through an uh in invocation of a web  service here on power admin desktop and that's   what i'm doing here so i'm collecting cpu  i'm collecting memory i'm collecting couple   of telemetry which i think for that specifically  for that specific flow uh would be valuable so   this pro actually is a demo flow which takes an  excel sheet with a list of countries and writes   this into another excel spreadsheet right so  this is a list of countries here very simple   and what i've decided here for that demo flow  that i would like the current context whenever   it's looping through that i'm always sending  to my power bi streaming data set the current   country so the current role in my excel i'm i'm  i'm working with okay so in order to show you that   um you can define a so-called dashboard power  bi dashboard which can host live connectivity   to those services sorry to those service execution  as you can see it has just changed because i have   submitted a couple of new flows right so if i'm  running those again here let me resubmit those okay so this will take a second until they warm  up but for that we have our nice view here you see   one is running one is uh queued up and the one the  other one is the next two to the right so if i go   back to my uh power bi uh live dashboard here as  the execution happens those records will will come   in and then you see here the last action which has  been called is right to actually you see now it's   in austria it's running now and you see if i want  to look at the at the the specific row i can see   here also the progress so it's live updating as i  said every couple of seconds you can define this   you see the memory consumption and also here the  cpu consumption on that and you see the expected   loop count of that process should be 196 loops  because that's the amount of countries i have   and i'm currently at position 80. so you know  you can judge okay it will be finished maybe   in two minutes or it might be finished in 10  hours right and this level of view where you   see of course also the cpu utilization uh a free  memory or whatever telemetry you would like to   to get from your power to make a desktop host you  can derive that here and then you can filter by uh   action by host again and so on okay that's the  and then you see of course here the development   over those machines i can look at the specific  machines and so on and from here again if i   would like to know okay what exactly that specific  flow was i can go here again to the flow history   and this is as it is executing showing we  then of course at which stage and position   we are in the cloud flow sequence okay so  this is rounding up actually uh john the uh   you know real-time processing with more uh you  know a a history processing of uh the analytical   uh data i love it okay so you see it's still  executing i have two buds to be finished and   this is what you can see here directly from  this dashboard as well okay any questions no   questions i'm i'm overwhelmed that's a lot of  information yeah and as i said no no c-sharp no   python code needed to to do these things it's  really configuration and the access to the cds   of course back-end that's cool yeah pranav  you've been quiet what do you have to say   i am mesmerized in real time i know that awesome this sort of fills up a  huge uh sort of gap that we had from a customer   perspective you know we have the out-of-the-box  analytics but that's sort of meant for a general   purpose audience and gives you a sort of  one lens into your desktop and cloud flows   but as you start to write more and more these  business critical automation on the platform and   your the level of sensitivity that you want tends  to be a lot more higher you want more more real   in real time both insights and auditing so this  gives you a nice sweet spot layering on top of the   out-of-the-box analytics all the data as possible  saying we're still in dataworks so it's the same   source that we're connecting to it's not that  we're copying the data over to some other data   source and now we have this data fragmentation you  know problem that you're running into it is just   a different view over the same data set you know  without having any c-sharp python or data science   skills so this is like beauty that's why i was  like uh totally mesmerized it changed real time   it's amazing love it yeah it's really great i  think uh you know i'm gonna ask the question   that i know everyone watching wants me to ask and  they're gonna ask how do i get my hands on this   and so for those watching who have been mesmerized  as well uh apostles how do how do others do this   yeah it's a super question i think yeah you  know the whole reason what i said before is to   give this extra level of control to flow and  desktop flow makers but also the ce folks and   citizen developers whoever would have access  eventually to that back-end data and that to   that level of data can of course make  use of those things and in order to   give that beauty to everybody that's why it is  called also starter kit and we are planning to   announce this soon and uh yeah provide this  downloadable component as a power bi template   all you need what i said before is  access to cds backend with sufficient   privileges and then having those two urls and off  you go this is for the desktop flow analytics in   order to build the real-time dashboard right this  is a different dashboard as you can see here and i   have i don't have it open that's a different one  because real-time data sets streaming data sets   cannot reside in the same power bi desktop file  as the other ones that's why there are two files   but i'm planning i can not say any dates i'm  planning to do because it's quite simple to do   i'm planning to do a blog post on how to  replicate that on your own right this real   time that well because it just takes what i said  before a schema you just these two actions here   to provide something to that power bi streaming  data set and then on the power bi service to   define that schemas with this field you have  seen and then you can monitor whatever you   like uh from an execution perspective yeah but  they will be very very soon to come yeah okay   all right so you heard it here take a look uh i'll  update this video when it releases so we're gonna   put this video out to show a preview of what's  to come but keep an eye on the description as   one place i'll go ahead and update it with the  links when it releases also keep an eye on the   power automate blog i assume that's where  the blog post will release right Apostolis   yup perfect okay so keep an eye on that as well if  you don't want to keep coming back to the video uh   the news will launch there so you heard it it's  coming soon you'll be able to download a template   where you can plug in your information and get  some of this goodness yourself and that to me   is the most exciting part about this whole thing  perfect okay so some final words uh john on   because the deck has been called  also behind the scenes or backstage   um so leveraging database this is how i've  realized this whole thing we have said that there   are three core entities you would like to look  into workflows flow sessions and file attachments   uh flow workflows uh this is where the desktop  flow uh are stored flow sessions is the execution   history and in a file attachment there's the  action context of those flow session executions   okay then i've used the web.contents for those  who know power bi because it allows you to handle  

different http statuses but also allow you to  include of course the odata annotations to get   from database not only the ids but also their  respective names for that then in order to   do an automatic or schedule an automatic  refresh on the power bi service   you have to use parameters with relative paths   and so on to to allow for that right it showcases  also some medium medium to to advanced data   processing commands in the m query language so  if we go back to our dashboard and we go here to   to edit the query just as an example many of  these things uh you know have been introduced   just to showcase of course so if i go here to  the actions and i look at the advanced editor   there are some things in there like if i'm create  if i'm getting the action context that includes   sensitive information like for instance a client  id or client secret or anything so i can really   obfuscate that information in the transformation  process and this is really showcased here with   a couple of those script commands you see here so  that's why it's also a good showcase on how to do   maybe power query and m transformations then this  is also very important data privacy fundamentals   so you cannot just say i'm ignoring the privacy  models because this is of course a security risk   but on top of that if you mix it up let you have  for instance for your connections set the privacy   on one data source to private and the other  one for instance organizational you will not be   able to uh publish so you will not be able to  automatically refresh that thing again on the   power bi service and there are some in terms of  where i said it's a good showcase to to do that   yeah you know and also to yeah to add to it that's  a very good point also because in your flows also   for folks who've been using sensitive text as  input output both for desktop flow and cloud flows   you know all that data is not logged  if you mark them as sensitive to those   that data does not get logged in in dataverse  it does not show up in those dashboards as well   so do sort of take a look at the uh the privacy  fundamentals don't take no there are a lot   of controls available both from a maker's  perspective and from a cv perspective on it   yeah absolutely true so whatever we see in the  portal you see also here of course eventually   right so some things like what i just described  with client id and client secret and so on so   those are things if you define them on the portal  and they're of course not hidden then you would   see them as well and this is where i'm doing  some minor checks on those but you have of course   to go into detail to your data and then make  your according your judgments whether something   is considered be sensitive or not and then  filter that out and so on so this is not a   fundamentals course on on security but uh just  to show you that there are means where you can   restrict that on on the power query and  processing level already yeah so then there's uh   the dax measures so the measures i've shown you  before so month over month changes and so on   and it leverages of course once it is deployed  to the cloud also the os awards to authentication   um yeah findings for those who are very well  versed in dataverse we have uh introduced   a concept which is called tds endpoint which  allows you to connect to a database environment   uh in to a read-only sql instance if you would  like right and this is of course very important   for for reporting so you're doing offloading  all your reporting load from the production uh   environments and from production execution and so  on and its respects again uh security roads and so   on so forth now the current implementation of tds  endpoints doesn't expose the flow session entities   that's why i had to use the web contents route uh  instead of the dataverse or common data service   standard connector we have there we are planning  of course to support this in the future but   there's currently no no eta for that uh okay so  it might be that the cloud flow history retention   differs from desktop flow history i think this has  been rectified already and for the large flow run   log history if you receive a 413 error this is  usually if you have super super massive large   uh legacy uh action flow runs uh which has been of  course resolved now in in the current version but   this is for the history you're having there if you  receive that then you know and the dashboard the   first page will surface uh those errors in the  action log here so it won't fail they will just   tell you okay we have five of them well i think  the the important part here is there is a lot   of power and you will be able to get your hands  on it pretty soon uh as we talked about before   keep an eye on the blog keep an eye on the  description of this video i will update it   when this comes out and you can get your hands on  it in the meantime if you have questions for the   team if you want to pick apostles brain about it  a little bit go ahead and leave us a comment or a   question down below we're happy to come back and  answer those as best we can all right this has   been it for episode number 13. you guys know what  to do go ahead and click like and get subscribed   so that you don't miss another video and that's  it from us we'll see you in the next one you

2021-02-18 04:20

Show Video

Other news