okay we're gonna go ahead and get started we are at the top of the hour uh welcome to platform analytics Academy um my name is Thomas Davis I am your host today we're going to be talking about enhanced data collector or uh optimized data collection um so it's a really great topic and some really good stuff that has come out with uh Tokyo so quickly I'll go over some logistical information and then we'll get directly into the uh content for today I can get this to move forward there we go so as usual uh this is for you so we're trying to bring fresh ideas and better understanding and some practical guidance it is being recorded and it will be out on the uh Community event page later on today and also out on the YouTube channel please feel free to ask as many questions as you would like inside of the Q a if we could put those questions in the Q a rather than the chat um that q a is being monitored by other panelists as well uh so please ask all the questions that you would like if there's a question that comes about that is relevant to where we are in the content I mean we may stop the presenter and ask that question then and see if we can get that answered if there are questions that are related to um something else that's unrelated to this content we'll try to get to those at uh the formal q a session at the end of this presentation so uh Safe Harbor notice uh please don't make any buying decisions on anything that is stating here although everything that you're going to hear should be um already released but there's a possibility that you could hear something so uh again please take notice of A Safe Harbor here and again my name is Thomas Davis and today I have lob and isoc that are going to be on and I believe that lava is going to speak first and izac is going to be there in support um and again we're going to talk about enhanced data collector and also have on some co-workers Adam Stout is here and Dan Kane are here as well and they're both going to monitor any of the Q a and also chime in um if they have something to share as well so with that I'm gonna pass it over to laugh so if you're ready you can definitely take control okay good morning good afternoon good evening everyone um my name is love and I am one of the product managers so I can get with performance analytics team and service now and the topic that I'll cover today is the optimize data collector that is a key capability that we delivered along with our Tokyo family release so the high level agenda is this I'll talk briefly about what optimized data collector is and what goals uh it kind of tries to achieve to give a quick or a better understanding I'll do a quick comparison on how the current flow of the data collector is and how it differs in the enhanced collection flow and then I'll get into some of the topics that are related to specifically to the optimized data collector uh and then a quick demo and then some benchmarks and key takeaways so what is optimized data collector um most of us are called today are aware that um you know data collector is the heart of performance analytics it is the engine that calculates all the scores and snapshots and that powers the kpis are key performance indicators that we track right so it is essentially an important or even probably the most important Cog in the overall Machinery that is performance analytics hence it is quite important that we ensure that we keep it well maintained optimize data collector is essentially a new implementation that aims to do that right and it aims to kind of make it more aware and considerate of instance performance parameters that is to say that it is optimized to efficiently utilize enough memory and CPU on the nodes and ensure that it does not overload them right especially when it is uh collecting for large data sets and that and that is another objective that it aims to kind of bring in ability to collect scores from very large data sets where score combinations are huge uh as the data volume of your clients grow and their needs increase this is this will help us ensure that we are ready to process more data and at the same time you know we do not cause performance issues for our customers so as I mentioned uh it is a new implementation or algorithm to collect scores so most of the changes related to this feature are backend algorithm and system design changes from UI perspective nothing much changes uh and I will touch up on that again later when I demo using an instance availability wise uh it's it's it will be there starting from Tokyo family release and by default it will be turned on which essentially means like all the jobs moment you upgrade to Tokyo uh all your jobs will run with this new optimized data collector we are currently also in works of bringing it to San Diego release tentatively uh and uh Safe Harbor here tentatively it would be uh around patch nine and on San Diego platform uh her family release version it would be inactive that means uh clients who need to kind of utilize this new optimized data collector they will have to uh create the system property and set it to true and this is controlled by System admins and here I provided a quick link to the overall product documentation related to this Optimus data collector so moving on um to understand well how this new data collector achieves optimization let's look at a quick comparison of overall flow of current and the new one I'll not go into technical details here but at a very high level the process comparison which will help us better understand how the new algorithm achieves performance optimization right the sort of flowchart on the right represents the current flow at high level uh scores are indicator Source uh imagine a job where there are multiple indicators the processing of scores happens on indicator source and for each indicator and breakdown combination we start by querying the most recent time period we calculate this course uh we then we calculate the score and then we keep keep them in the node main memory then we continue processing for the next time period and then for the next indicator Source until we are done and then at this point of time is when these scores are inserted into the main L1 error L2 scores table right so just uh as you can imagine in this flow when the job is like processing large data sets are essentially large number of scores are being calculated it can be highly dense small to medium data sets or even a sparse but large data set or any combination of it but ultimately it is the number of unique scores and snapshot that determine the memory utilization there is a potential that you know um the job will run into issues like low memory which leads to Performance degradation or sometimes may even lead to restart of the node and because the data collector jobs are not the only thing that fights for node resources there are other Mission critical production processes that also need those resources which can further aggregate the problem so how does the enhanced or optimized data collector is solve this problem we introduced a subflow uh into the main flow which is executed within the main flow and this is this essentially is the representation of the new subflow uh essentially the overall process remains unchanged but instead of holding all the scores and snapshots in memory the new data collector processes them in batches so the flow starts normally this course calculated for indicator and the recent time period but the algorithm is constantly checking if there is an optimal threshold that is reached and I'll talk about the threshold in the later slides when it reaches it it essentially dumps uh everything that intermediate scores since after that it has into a temporary location on disk uh and then continuous processing the new sub process will process this course for each combination of indicator and time period uh Aggregate and insert them from temp DB to the main tables and then it deletes the file on the disk so I know it's it's a bit too much process but essentially to summarize the new data collector achieves optimization by keeping only subset of intermediate scores and memory and dumping them into temp location on disk while it continues processing and in this way it limits the memory that it takes up I'm going to pause here and see if there are questions or comments at this point of time now we're good okay cool so moving on a bit on the threshold limits although they are a bit more of technical implementation detail but I'll quickly give an overview of a three percent just for understanding as we discussed earlier they decide when these scores and snapshots are offloaded to disk and it is a combination of three numbers as shown in the table here the number of uuids which is the distinct score combinations uh for the kpi for the indicator and uh breakdown combinations which is set to 100K number of snapshots and number of distinct value which is a property used specifically when aggregate is Count distinct if any one of them reaches we offload and we capture the information in job logs and I'll show that when I demo uh how it is captured in the job log uh what they represent is not important as such but because they are hard-coded limits and cannot be modified but the point I guess what we want to drive here is that these limits are result of uh quite extensive and structured experimentation on various data volumes and were decided as convenient trade-off between performance of both runtime and memory right when we optimize for memory we also do not want to affect the runtime of the job and I'll talk about that later how we actually improve the runtime in most of the cases with this new data collector moving on the data get collector limit properties so as we all we are all aware that there are limit properties today that exist uh which which specifically exists to kind of avoid such performance issues right with this new data collector most of those remain unchanged but instead we do introduce these three new properties which are counterpart of um the old data collector and I'll just quickly show how that is essentially the indicator row Source row count which in the old data collector is set to 50 000 uh in the optimize data collector it's set to by default to 1 million and that means it it can process a quite High number of uh indicator Source records and not run into performance issues similarly with maximum breakdown elements and for L2 level 2 we have these limit properties which are set today to 10K and 1 million uh there these new properties which apply to the optimized data collector essentially have no limit right so by default no limit is set on these but If you experience these properties exist because if you experience performance issues and we do recommend testing uh in subprod instances before you start using it in fraud uh you can set these limits to whatever optimal you find it but we uh so the key Point here is the default is higher for the optimized limit and this this is something that we have identified internally based on our testing and uh benchmarking so moving on uh job logs so when it comes to the overall data collected job logs most of it is same in fact 95 of it is same there's just additional logging that we do uh when the data collector is running with optimized data collector so uh the two of them which I want to call out here is one uh the first log that is captured is also captures which data collector is it running on so here you see in the screen it's saying using hsql data collected true which is the optimize data collector uh if it was old you would have seen like a nested collection or else and the next one here is we talked about offloading uh based on the threshold hits so that is also something that we capture in the logs and there were three limits that determine offloads and all of these three are essentially captured separately so here in the screenshot it shows that you know because the score limit hit twice it had to offload and I'll walk through this again uh during the demo so I'll pause here and see if there are questions before I move on to demo uh there there is a question and isoc is is actually typing it um but it may not be bad idea izak to actually come off and uh love you can as well yeah I was gonna suggest that as well so esoc if you're answering for Lisa if you could do that one out loud as well yeah I'll take it like with will this work with current PA jobs uh or is it something we have to transition our jobs into uh it will work so uh I think that is what I was trying to drive um on the first slide essentially you don't have to do anything uh both of these Logics exist so moment you upgrade your instance to Tokyo um it all your jobs will essentially just run on the new data collector uh you don't have to do any transition test so declare it's really just a question of whether the new engine is on or off yes the for the entire for the entire instance for the for PA yeah okay that's good that's good thanks and I see there's also another one that says uh when moving from San Diego to Tokyo will the properties be fixed upgrading to Tokyo or will the customers need to manually change them so when upgrading the token the property uh by default will be true and in case you want to turn it off you will have to uh as I put it in the first slide you will have to set the proper to create the property and then turn it false uh but when it when you are just upgrading to San Diego patch 9 there is an action required essentially There It Is by default false and you need to kind of turn it on a question from Raymond is San Diego patch 9 is due about November 1st I think it is uh December 1st uh as per my knowledge goes and question from Stephen is for legacy customers will this property be enabled by default on an upgrade right yes if you upgrade to 2q it will be on by default the use case I can think of within that too lab I just want to make sure so if I had let's say you know I have set the global property already up to let's say 250 000. um in the the Legacy system and I've had several individual indicator sources that I've overridden it to be like you know a million or two million what will be the net impact on the upgrade the the uh the default Global one will just automatically be a million regardless is my understanding right right um so there's there's a check that's a good question there's a check that we do uh when we run the script essentially if uh these are the defaults for the new one um but if if we see that uh the old defaults are higher than this then we actually take those values foreign okay so in my scenario my example if I had set the default to 250 000 it would remain at 250 000 for the new property as well okay now for the new it will be million that's okay that that's what I'm saying here yeah that's what I'm saying but if it is higher than million for old uh then it will be uh higher than million for the new gotcha okay and then for any individually overridden I think that works however it works in the property that at the uh you're talking about this over Source level right that was the main thing those remain same those don't we don't change that at all no we don't I see two more questions uh apologies if I missed earlier Josh is asking if the snapshot limit increase uh at this point of time no so snapshot limit Still Remains at 5K moving on um I'll quickly get into an instance and show as I was earlier talking um there isn't much difference from the UI perspective but I'll just quickly talk about what all things changed so moment you upgrade to Tokyo as I said earlier this isn't much much of the change to UI for those upgrade in Tokyo there's no action required the moment you upgrade your production instances to Tokyo the job will start running using the new optimized data collector and if you want to turn it off uh you essentially create this property here and set it to false right but those who are on San Diego uh you will need to upgrade to patch and you upgrade to patch 9 you will need to manually turn this property on to utilize the new data collector so when this property is turned false it essentially continues to run on the existing or the old data collector the jobs will continue to run on that and when it is set to true it runs with the new optimized data collector so um let me take an example of a job uh before the session I actually triggered one of a job and I wanted to essentially show a comparison between the old and the new so here you see I've run this job twice today this one it essentially uh inserts 168 000 scores in snapshots uh this one is this one is running on existing data collector you can see here is nested collection true and it took around 12 minutes uh to collect all this course and complete when compared with the existing uh the optimized data collector which is you can see here from hsql data collector true uh it essentially completed the same job in two minutes which is like a huge uh runtime Improvement um and that is something that we have observed in our extensive benchmarking and shared the benchmarking numbers later uh that on an average case you you should see uh your jobs running a bit faster uh maybe 50 to 100 percent faster in many cases or at least they will be comparable runtime they will never be um slower than the existing one and at the same time they're memory conscious so just quickly I also showed you the logs earlier on the slide so this is the one that kind of differentiates whether it is using the new optimized data collector or the old data collector and I also did talk about uh the threshold limits so here in this job you see it it captures three logs for score snapshot and then also by counter strength here so essentially this job offloaded twice based on scores so those are the additional logging that we do for this job uh apart but other than that most of the logging is quite similar uh with whatever is with the old data collector and this is done intentionally uh so that you know going back to Lisa's point you don't have to transition over too much and it works like and seamless experience after box and the last thing I we did talk about [Music] um limit properties so I do want to show them that if you go to the performance analytics property page and these properties are categorized by their category you go down the page performance analytics data collector you'll see these three new properties essentially which will control the indicator Source records or the L1 or L2 breakdown Elements which you can utilize our recommendation essentially is that these are the default values one of the questions that we usually get is can we increase them um I think um the response that we want to give there is that you know it entirely depends on your testing right these are the default and maybe if you want to try you can just answer that we got the two questions that are out there right now I will go through the questions yes thank you so the first one is uh if the new data collect if well the optimized data collector takes the data outside of our servicenow instance and process them in another server and then move the data back and this is not the case so our data collector just we just changed the algorithm behind it the data remain on the server and on the on your instance we don't take the data out of your instance everything remains in the instance and they don't move to any other server so there is data always stays in our uh in our database server and in the nodes that you currently have sorry my network drop am I audible now yes okay I see you just answered the Lisa's question he said so yes the data doesn't move out of instance it just moves to the disk and then back to the table that's it uh this question from Stephen was customers intending to skip the token release going straight to Utah what service now recommend that we enable this manually in San Diego past 9 or whatever the official release winds up being if if if that is a need Stephen like if you see there is a need for processing large scale data or you see performance issues then yes you can you know enable this because this is exactly what it was built to kind of handle right I think that would be a short answer and I'm wondering maybe this is part of what you know if anyone else has the thought too is there any disadvantage if I go to as if I'm at SD patch 9 or again as Steven says whichever ends up being this the formal the official patch that supports it is there any downside for me turning that on other than just you know not being aware of changed there shouldn't be released there's no downside yeah so Stephen just I think I understand where you were coming from so and the reason why we enabled it out of the box in Tokyo so whenever you upgrade you get it enabled is that purpose we are confident enough to uh to move you to this data collector the property is there for that purpose just in case there is something that we might have missed and you can always move back to the old data collector but the idea is that everyone uses the new one okay Stephen came back with is there a downside no there is no downside I think you get all the benefits from what we learned throughout the years from the data collection and that's why the properties are not there and as you can see we removed a lot of the limitations we currently have and uh the I think the benefits of the data collection is something that love is uh going to come to right after I think yes yes um yeah there's no downside as such I think maybe the confusion is from why it is not by default active on San Diego is essentially because San Diego is out right and this is one of this is a big enhancement that we've done part of Tokyo but then we see that there are some customers on San Diego that can utilize it right right um so we are essentially kind of bringing it now to San Diego I guess that that kind of makes it a bit more conservative for us and we kind of kept it by default off uh but you also do remember Stephen that you know there's a so but yeah so I think it makes sense what what lab say in there on the uh um the the really the main reason that we're not doing it by default in San Diego is just to be consistent in that we we don't like to turn on new features by default especially in in uh in a release that in a patch or at least already out um so it's more of just being consistent with that correct and also the reason why because this is more or less it although it we replaced a feature that is already there the data collection it's quite big and usually we don't back Port features in previous releases but we did this because we know what is the benefit that you'll all get from this and that's why we back parted it but we kept it disabled with the option of you enabling it after testing which lab is going to come to uh I'm sorry I again dropped off I'm I'm actually sitting in office and somehow the network is right um okay um so quickly going into performance benchmarks overall as I said earlier in most cases the scores are calculated up to 50 to 100 faster so these represent some subset of jobs that we did internal testing on uh although there were like way more jobs more than like 100 or 200 jobs that we actually tested on uh with different configurations but essentially the point you want to drive is that uh these are some decent volume and um some decent number of scores that are being generated here look at this job which essentially 10 million records or 28 million record is trying to process and generating 6.3 million score combinations right and it did not complete uh with all data collector we know that it hits the limit somewhere because it runs out of memory uh with the optimized data character however it is able to collect it right and if you see in most of the cases here uh these jobs on time is quite faster which essentially means you can accommodate more jobs and uh there's less need to kind of split jobs I mean however it means to the outcome to you so we are quite confident on the solution as such um and we only see advantages positives with this so moving forward uh key takeaways that we want to kind of uh take from the presentation today a primary outcome with this new optimized data collector is to optimize memory and CPU consumption of the nodes and avoid causing performance issues while at the same time it also provides comparable runtime or even better run time in most of the cases right so scalability at least with our internal testing uh we see 10 x Improvement in the overall data volume handling right um you can even process more than that if you are okay with jobs running multiple days but but it's never gonna be kind of failing smaller job memory footprints that does not grow on the data scale which is essentially we see at least 4X Improvement in the node memory utilization so if earlier a bigger job would take like a GB it would not Maximum take 250 300 MPS of memory and jobs will run 50 to 100 time faster which is 2x Improvement it's available by default on on Tokyo default of on San Diego ps9 you need to create this property and switch it on or off however it works for you it has higher default collection limits because it is scalable it takes less memory and can process large data and one key point which we want to kind of point out here is that you know we still like we recommend that you consider testing on some broad instances to identify the right limits for you uh even though we have default higher collection limits uh and some couple of properties we have no limits but if um do you see some issues if you want to set some limits you're essentially uh able to do that before you enable it in production um that's all that's all I had for today um I'll take any questions or comments that you have in this point of time awesome thank you uh really good stuff and and stuff that I know that people have been talking about for a long time being able to increase and and things as it relates to data so uh really awesome I appreciate you sharing that information and sharing with anybody else so um Isaac do you have anything that you you want to add before we open up to um general questions no I think I'm I'm good thank you Thomas yeah okay um Dan Adam anything awesome stuff I agree yeah okay so if you we don't have any outstanding questions at this point so and and we still have about 26 minutes so here's an opportunity to ask uh other questions if you have any of those so um if you do please ask away now I know that there's like 12 people that are just typing away so they should pop up any second here foreign everybody's going to go turn on this property right now so yeah that's what's happening testing it well maybe we don't there we go review the new dashboard in San Diego should we put Raymond on uh allow him to speak and ask so we can get more clear sure there you go Raymond you are hey uh I was just like the difference between the guided version and the uh more traditional version of dashboard and um and I'm but the guided is like the workflow like where you're almost doing workspace where you're creating reports that are really data objects as opposed to putting a report on a dashboard at least when I was when I was testing San Diego that's what I was seeing right are you talking about like uh putting a the in the um in the the workspace or next UI type the visualization next to I visualization okay yeah so within San Diego are you using them still the the UI Builder as the basis that you're talking about here we have now we're just we're just rolling into San Diego so I'm testing and I'm looking and okay I saw that they have the new interface but all I could find was the more complex version of building the dashboard dashboard in the next UI there was there was one that's more drag and drop from reports that you already have on this one where you create the objects on the page and I was having some sizing issues it's possible that um and anyone else will feel free to chime in here too but we had a an interim version of the I'll call more of the simplified dashboard builder in earlier versions of the next experience UI actually it was called dashboard Builder there's the UI Builder and then dashboard Builder which was just kind of a stripped down version of UI Builder dashboard Builder is being deprecated and being replaced by a platform analytics workspace starting in Tokyo yeah what you're looking for is what's in Tokyo which I think we did a session on it which I'd look at with the inline Builder that allows you to build data visualizations like we used to build reports but data visualizations whether it's PA or reporting you build those if you want and then put them on New dashboards in the next experience or you can build them on the Fly we really get the most best of both worlds where it can be a one-off or it can be a reusable object what you know what works best for you well and I I like the one-off thing because that kind of gets that out of the I built a report now everybody can see it in recording but it's made specifically for how you have it on the dashboard so so now with in Tokyo where we're able to build on the Fly which is useful a lot we can build reusable when we want to which is and they're the exact same right they're the exact same just do you want to make it reusable or not ostensibly and we layer that with um report sources is what we used to call them they're still report sources in next experience their their labeled predefined filters I think we layer those things in and we can do some amazing stuff just the way we want right we want to make sure the data is consistent we use a predefined filter we want to make sure this report gets reused we do we do that um so lots of great options you really life's easier in Tokyo um are all report sources available to everybody automatically or is that like can you control that I don't think we have role based control on those okay um I can see that you can just put a traditional filter on it without doing a a source oh yeah you don't so so the way that it's structured before um and and the Legacy report Builder you'd pick a report Source right do I want a table or a report source in next experience you you pick the table and there's just free to find filters so you you put in the table and you get what you want um you see those filters they're just can filters um right but you can still customize it but then you add more right and which you were able to do with report Source before um you don't have to use a report Source at all it's just there are certain times where you know you need these three three conditions to get the right answer yeah you can save that um yourself on a one-off that's fine we're working towards that but we're struggling because they keep changing the the um when I want to say the ground underneath my feet with VR well and and that's where you can get um when we have the same table you can you can have the filter and you can just change the filter right change it yeah specifically yeah I was working on a dashboard last week and we were refining the query but we had five or six reports that were all versions of it just similar to how we have an indicator source and indicators it's the same way I think about report sources and and filters you don't have to put a filter on it all but it's useful to start with that and then I just have my slices of that they're adding additional variations to it but particularly when my base data set is changing conditions wow it is awesome to have a predefined filter yeah okay there's lots of time and again you can use that on a one-off visualization or a canned visualization so you you have a lot more um uh levers to pull I I think the the way that I summarize it is the same thing that um uh Isaac and live were talking about for the data collector we've been listening to you some of you have been telling me the same thing for many years we do listen and we're you know making it better and going well this is what you really want to do so let's make it easier to do what you want to do and and that means sometimes we get rid of features that nobody uses because it just complicates things and we simplify it in other cases we're adding new options to give you those tiers and I think in next experience um you know I know that everybody loved having PA visualizations be different than reporting no one ever wanted that um now they're not right they're not it's the one-offs the the cans sometimes you you sometimes you feel like a nut sometimes you don't um you get all those choices in San Diego you know don't go ahead and look and you have uib and uib it unlocks everything you could do everything in that I don't use uib very often I'm using the inline editor I went to go redo a dashboard and it took me a couple hours to see some of the options and then I'm not going back once you get a taste of the inline dashboards the in my dashboard editor in in Tokyo you're not going back okay it's awesome Yep yeah well and what I didn't see with the the dashboard Builder was like I just wanted to pull up a report I already had so super important so in next experience right now the old reports do not come in to next experience um that that is absolutely um something we are we're working on we are looking at how do we make that happen um to to come through I know that that that that is um you know we're not going to be you're not gonna be able to switch over until we bring over the stuff you've been working on for the last couple years right a few years several years so that's something that we have to have in there uh uh to make end that is not in Tokyo um today if you have a report that you should have on a responsive dashboard that's a report you cannot access this Deja visualization you need to you need to recreate it um my suggestion for for all of us is go look at the news in Tokyo go look at the new stuff start looking at the data visualizations if I have a net new project a net new dashboard go create it the next experience um if that's something you're gonna do you want to make sure you're looking at that you're considering that seeing what's going on if I if I'm just tweaking a report that's out there in a dashboard everybody uses just tweak the report on the dashboard that everybody uses um we'll we will figure out how do we help you migrate to the new stuff okay other questions but they're very specific to what I'm working on so when you're done if you got a minute okay um and and as you have those questions you know make sure that we'll talk about what we have here um the community is a great source to do that as you're looking through there's there's the analytics uh community that we're all familiar with there's also the next experience Center of expertise has a lot of great um articles and and some questions that have come up about how do I use uib and get the most out of next experience my goal is that you don't have to go over there but if you are trying to do some Advanced things you can in the past with reporting and Pa and dashboarding you're either out of the box and someone Limited or you went off you went into a major customization land with custom content blocks and macros and those of you've done my lab on custom interactive filters you could do a lot but it was hard I mean it was just hard in the new world we have inline dashboard editor which gives us a lot of power and we can do a lot of things um staying completely in the box and then uib customization is now something we configuration well it used to be customization is now configuration with uib it can still be challenging but you can also do amazing amazing things when our um in the past I mean I I know a lot of us have had this where I had a cxo who just wanted it to look a very specific way and and that's not how the PA widget worked okay you know there's nothing we can do well now with uib you can make it to look how you want it to work look work how you want it to work um again you want to make sure that your um you can do that for most of us that are on this call that are talking about what we normally do the inline dashboard editor is going to just answer a lot of the concerns that we've had um there is another question about um permissions and and the next experience um if somebody can correct me but I as far as I know I and I most of the instances that I work on people give me extra access so I I could be wrong here but I believe the dashboard the analytics workspace the analytics workspace is available to anybody once you upgrade to Tokyo yeah that is correct Adam the the only other required role that you'll need is if you wanted to save something to the library then you would have to have the createviz uh role but by default yes you can perfect right so it's there uh it's there it's awesome um again whether or not you're switching over to it all the way you know we'll help you get there um hopefully it's not going to be five or six years like for us getting rid of home pages um but absolutely we understand we need to help you get there I recommend you watch the previous Academy sessions that we've done on the next experience the inline dashboard the whole recent um the what's new in Tokyo and I think we have some sessions coming up on those things it's amazing um I don't want to ruin the Thunder but I'm not going back um I'm next experience all the way uh but make sure you're getting familiar with it so that when we are when you are ready to move over you're not having to learn the new tool as well um so whether you're doing it on your instance or you get a PDI then go take a look at it we'll get some Labs out there I'm sure we'll be tons of stuff for knowledge next year hopefully some more before just again very much like the the data collector we just talked about where well why did you want that limit nobody wants that limit let's get rid of the limit I think that's what we're going to see with next experience and the inline dashboards a lot of limitations we've lived with just go away yeah good stuff thank you Adam so I don't I don't see any more questions that have come in so I'll go ahead and get through these final slides and if another question does happen to come in then we'll get on that so okay so as usual make sure that you're checking out the community like Adam was referring to earlier if you have uh questions or lots of questions make sure you check the Community First a pretty good chance it's already been asked if it is not you can definitely ask it there uh not only do servicenow Personnel uh answer the questions but we have plenty of customer gurus that are out there that know quite a bit and answer a lot of questions throughout the day so make sure you're checking that out um this again will be out um on the community later on the recording of this on the platform analytics Academy Community page and also um any of the existing things that we have had and things that we also have coming up um something new that we have launched is platform analytics Academy Pro tips and what we're trying to do with this is uh we're doing little Quick videos so four or five six minute videos where we actually talk about a specific topic and we expect this library to grow um quite a bit and our first uh series is where we talk about you know uh first value with performance analytics so if you get a chance go out there and check that out and of course all of this is on YouTube as well and I never forget about now learning a really good training that is in there and uh new stuff gets released and launched all the time um you know the best way to find anything that you may need as a release analytics is just by simply typing in analytics and all of those things will come up uh and then the PA Advanced uh I've said this for the past few sessions it is now on demand so if that's something that you want to do on demand uh you should definitely take advantage of that these are all the other uh sessions that we like to plug uh really great content on all of these just like our academies they do some really great things so if any of these topics that you're interested in I would definitely recommend that you get over to those and check those out and in two weeks uh we'll have our next performance analytics Academy open question and answer session these are really exciting and uh generally draw a lot of traffic so here is an opportunity for you to ask questions to the professionals we will have a panel of roughly about around seven people and uh give or take so again this is a great opportunity for you to ask all the questions that um that you have and um you know get some really good information from uh professionals that potentially develop this and also work in it every day so look forward to seeing everybody in in two weeks and it looks like that we don't have any other questions so looks like that we'll get some minutes back today about 10 minutes back so with that thank you as usual and we'll see everyone in two weeks uh thank you to the presenters and all of the panelists for coming and supporting we really appreciate that so we'll see everybody in two weeks thanks