Observability for the AI Era: Governing Data, Cost, and Risk with Governance Hub

Summary

Governance Hub is a unified account-wide observability center in Databricks that provides a single interface for monitoring governance changes, identifying risks, and acting on cost and compliance issues across multi-cloud and multi-region environments.
A built-in recommendations engine automatically surfaces risks — including runaway AI spending and access policy gaps — with one-click remediation so teams can resolve issues without navigating multiple tools.
Natural language governance queries powered by Genie allow teams to interrogate their governance posture conversationally, while system tables expose the underlying data for custom automation workflows.

Observability for the AI Era: Governing Data, Cost, and Risk with Governance Hub

As data and AI estates scale, account-wide observability becomes critical to maintaining control, trust, and efficiency. In this video, we’ll introduce the Governance Hub, your unified center for governance and observability in Databricks. You’ll learn how to monitor changes across your environment, identify risks, and take action across data and cost governance from a single interface, with expanding support for areas like performance and AI governance. We’ll also show how built-in recommendations help you surface and resolve issues, and how system tables support programmatic governance and automation workflows.
Talk By: Kristen Wilder, Senior Product Manager, Databricks ; Michael Postiglione, Product Manager, Databricks ;

Chapters

FAQs

What is Governance Hub in Databricks?

Governance Hub is a unified account-wide observability center on the Databricks Data and AI platform that consolidates data governance monitoring, cost oversight, and risk management in a single interface. It supports multi-cloud and multi-region environments, giving platform administrators a complete view of their governance posture without switching between tools.

How does Governance Hub help control AI spending?

Governance Hub includes cost governance capabilities that track AI resource usage in real time, enabling teams to catch runaway spending before it exhausts budgets — as illustrated by the well-known example of a company burning through its entire AI budget in four months. The recommendations engine surfaces cost anomalies automatically so teams can act proactively.

What is the recommendations engine in Governance Hub?

The recommendations engine automatically identifies risks across the account — including access policy gaps, sensitive data exposures, and cost anomalies — and presents them in a Findings view. Many findings support one-click remediation, allowing administrators to resolve issues without requiring a full manual investigation each time.

Can I query my Databricks governance data with natural language?

Yes — Governance Hub integrates with Genie so teams can ask natural language questions about their governance posture, access patterns, and compliance state. System tables also expose the underlying governance data programmatically for teams that want to build custom automation workflows or integrate governance signals into external dashboards.

Full transcript

[00:08] Um okay, just a silly slide, but we know it's after 5:00 p.m. So, thank you guys for spending your happy hour with us. We're excited to talk about governance and observability, and hopefully we can get you some beers after. Um okay, going to flash a forward-looking statement for all of
[00:21] you. I'm sure you've seen this maybe like a thousand times already, but um just want to caveat that everything that we're showing today is representative of direction and roadmap. Uh that being said, we're not committing to anything in this presentation at this
[00:34] time. Um okay, cool. So, let's talk through agenda. First, we're going to talk about what makes governance and observability hard, where the challenges we're running into and seeing amongst all of you.
[00:46] Uh next, we're going to talk through our new feature governance hub. We're going to talk about recommendations. We're going to do a quick demo. And then we're going to talk about where we're headed next. And we're really excited to talk about that last bit.
[00:59] Um okay, so this might look familiar to some of you, but there was a meme going around from this guy named Benjamin uh a couple weeks ago. So, I decided to screenshot it. Uh he he was doing some whole post about Databricks, Alation,
[01:13] all these different vendors and choices, but the part that actually resonated with me was there's this guy, Bob. I did make him look a little bit more special and put him in purple. Uh but there's this guy, Bob, who's responsible for governance across the
[01:26] platform, has all these decisions about vendors, governance, etc. all falling on his shoulders. Um so, I thought it'd be fun to include it, and I'm actually curious. Um everyone raise your hand if you've
[01:38] ever related to being the Bob of your organization. Yeah, that sounds about right. Uh well, you're probably not surprised, but you're certainly not alone. Uh here's a couple news articles that we've
[01:51] been seeing. I think my favorite one of these is the Uber one cuz it's so relevant uh where they ran through their entire AI budget in the matter of 4 months. Uh and I think it's really telling. Like it shows the importance of
[02:04] having proper budget set up. It shows the importance of having true observability so you can actually see how people are using AI and see some of these things ahead of time so you're not being reactionary.
[02:17] Uh I'll talk through in an ideal scenario, governance is kind of a continuous cycle. Uh so we'll bear with me. We know it's not always perfect. But first, usually you can start with definition phase. You
[02:30] have a set of policies or standards that you want to define. And then usually you need to go into maintenance mode to ensure that the policies stay enforced, whether they be budget related, access
[02:42] related, what have you. Next, you usually have to go monitor changes over time cuz you bring new data and you bring new AI sources in, you bring in new users. And then you usually need to go through a detection phase where you
[02:55] can spot anomalies and identify risks as they come up. And then oftentimes when risks come up, you have to respond to them. And that might look like adjusting your policies and you know, redefining the
[03:08] policies. So that's the ideal case. We know even internally that that does not always happen. So here we have Bob. He is laying down cuz he's sad and I couldn't find another
[03:20] way to make him look sad. Uh but I wanted to run through a couple of different problems that we've heard from all of you and other customers that we've been working with and partnering with on this project. So the first one
[03:32] is spend limits. Sometimes they're tricky cuz they have to be manually set across different workspaces and there's not like one spot in the UI to go manage them across your entire account. Just
[03:44] one example. Another one that's particularly important to me just cuz I yeah, you know I'm partial about tags is maintaining my data super slow and manual. And then in terms of monitoring, again, kind of calling back to that Uber use
[03:57] case, uh it's really hard to identify who is spending the most and who's actually using AI within your organization. And this is becoming more and more important. And then in terms of a detection
[04:10] scenario, so a lot of times we hear people are hearing about escalations and anomalies through their teams. Like rather than being proactively alerted, they have to find out via Slack, which is not ideal.
[04:23] And then finally, I'm sure some of you guys know this. Once you've seen the the anomalies or have detected these escalations, often times you have to go and fix them one by one. Uh which can be really slow and again,
[04:35] manual. You're probably noticing a theme here. Um so that's where Governance Hub comes in. I'm going to focus this first part on these three buckets. So define, maintain, and monitor. Uh this slide is a lot uh and I will
[04:49] break it down. Basically, we're trying to focus on creating a single pane of glass for governance across a number of different verticals. Uh we heard a lot of feedback through various dashboarding tools that we've created and given to
[05:03] you guys to test over the years, and we got a lot of feedback around wanting really that like account-wide visibility rather than things being scoped to the workspace. And we also heard a lot of complaints about, you know, the slowness of
[05:15] querying SQL um and system tables directly. So we decided to build Governance Hub uh and prioritize speed here. Uh so right now we're focused on building out governance oversight across
[05:27] a number of different verticals. Uh the first ones that we've kind of targeted are data, AI, cost, and performance. Uh in addition to be able to see these account-level insights, you can also drill into specific workspaces,
[05:41] metastores, and then also, a new one is clouds. So I'm not sure if you guys were able to catch Z Sean's session uh earlier today, but we have a new multi-cloud governance
[05:53] strategy where you can actually have multiple clouds per an account. So, I'll give you an example. Like let's say you have some data in AWS, you can have that data be in an AWS workspace, and then
[06:05] similarly you might have data in Azure. Rather than having those be completely separate accounts, we're finally solving that problem and you can have them all in one uh account. And what that means is with governance hub, since it's account level, you can govern it all in
[06:19] one UI. Uh there were a lot of conversations about Genie this morning, so this probably comes as no surprise, but we are also integrating with Genie to ensure that you can ask follow-up
[06:31] questions about uh the insights that you're seeing in governance hub after that. So, I picked on three different use cases. I figured this might be a little bit more fun than going through like a
[06:43] list of features that we're supporting uh cuz we'll do a demo and show you those after. But, calling back to the metadata maintenance problem, uh one thing that I hear a lot is that
[06:55] it can be kind of tricky to figure out when you have a lot of data, it's all untagged, how do I identify which ones to prioritize first? Uh like how do you actually clean up and find a needle in the haystack in terms of what you want?
[07:07] Uh so that's kind of why we've introduced this new usage view. Uh essentially you can actually identify which tables and other assets are most frequently used within your environment,
[07:19] so you know which ones need to be prioritized first. So, for example, here in this view I see Lee has a whole bunch of tables that are frequently used, uh but they have no tags, quality isn't enabled, and there's no comments. So, I
[07:32] might want to go bug him after this. Uh and here's another example for cost. So, one thing that's come up quite a bit from our the our cost PMs hear all the time is it's really tricky to attribute
[07:45] cost per a product line. So we actually introduced a new cost explorer which I'm super excited about where you can filter and view based on like model serving all purpose and then also break down by things like workspace
[07:58] and tags. And then finally for the the AI example um back to that same problem of understanding what which users are actually using AI, how much they're
[08:10] spending, how can you enforce budgets if you don't know what they're spending. We're introducing a new AI usage view that's built on top of Unity Gateway where you can actually see which users and groups within your organization are
[08:23] spending the most and using AI tools most. All right, I'm handing over to Michael for recommendations. >> Cool. Thanks Kristen. Um yeah, so we
[08:35] Kristen just walked us through how we have a single pane of glass now uh in the product of the account level uh to better monitor uh all the different aspects of the Databricks platform. Now we're going to shift focus talking a little bit about recommendations uh and we're going to
[08:47] focus on how you can really uh both detect and respond to incidents that happen across your Databricks uh deployment uh at scale. Um so today right now spotting anomalies
[08:59] risks as we as Kristen mentioned mentioned uh is largely a manual process um happens often being a slack not proactively. Um so we want to fix that and be able to automatically surface things to you and allow you to take action uh at scale
[09:12] once you identify those. So that's where recommendations come in. So really the the impetus for this problem uh is that you know we have all the BOBs uh here right here uh and basically everyone owns the governance and
[09:25] actually doing this at the scale of your enterprises just doesn't work when you do the math. Right? So you have for example 12,000 different assets uh there's too many to wash. Every single week they're going up, they're growing, they're not getting less. On
[09:38] every single asset you have, you know, 10 or more people who can actually change the configurations and alter things without you realizing. And by the time you realize you have a bunch of Jira tickets and escalations cuz your
[09:50] stakeholders have found out. Uh it's just too late. So, to actually solve this issue, we need to really introduce automation into loop. So, we need to automatically detect when changes occur to your different assets
[10:02] and resources across Databricks, surface that proactively, and then give you the tools to respond at scale to get to these problems before before everyone else realizes.
[10:17] So, this is recommendations. Um so, in the governance hub, we're going to be introducing automatic recommendations, which are your best practices across Databricks. So, we'll have them for all the different domains, including data, cost, AI, and
[10:29] performance, and they'll be surfaced directly in the governance hub on all the different pages. Uh they'll be available in my handrail. You'll be able to see what the best practices are, and we'll automatically flag like the top assets and resources
[10:41] for that particular recommendation and give you additional screen to click into to review further and understand what you actually want to immediately triage or or show to other people.
[10:53] So, with that, let's kind of go into end-to-end workflow on one of these recommendations. So, what we're looking at here is a cost recommendation that's in the governance hub. Uh and it is for cost attribution. So, in Databricks, to
[11:07] do cost attribution, you apply tags to your different compute resources. Um and at the end of the day, if you don't have things that are tagged, it's basically invisible when you are splitting the the bill across your different teams. Um
[11:19] so, this recommendation actually goes across all the different compute resources you have in all your workspaces, surfaces everything at the account level, um and shows you your highest cost untagged resources. Uh this is really the detect step uh in in
[11:32] recommendations. It allows you to to see everything. Then once uh you're at that step, you're able to actually click into the review page and get more details on all the different findings we have uh across
[11:44] your account. Uh so you'll be able to see what type of resource it is in the case of this recommendation, who owns it, and then how much it's actually spending over the last 30 days so you can understand its uh importance. You can investigate, determine what you
[11:56] want to act on, and then even click into the sidebar to get more details, and then go straight to the workspace to to deep dive in there further if you want. So the idea is there's no hunting for context across multiple workspaces. Um
[12:08] but you can just dive in and understand uh immediately without without switching screens. And then lastly uh you're able to actually go and resolve these fixes. Uh so today you're able to click indirectly
[12:21] to the workspace, go to the cluster uh page immediately. Uh and then actually just go and apply the tag uh directly to the cluster. Um so right now you're still in the driver's seat applying these fixes. Um
[12:33] but we'll touch on later about how this is going to change in the future with with bulk action. So with that, I'll hand over to Kristen to start a demo of the governance hub. >> Okay, perfect. Um do you have a clicker, by the way?
[12:46] >> Yeah. >> Perfect. I actually don't need it right now, so I'll just leave it here. Okay, so we're going to do a live demo uh and kind of just walk through what we have available right now. There are some new additions that even I saw, including
[12:58] performance, so I was kind of excited to see that shipped in our our staging environment today. Uh but in terms of what you're looking at, this is actually the account console. So I mentioned before how this is an account level
[13:11] view, uh and this is going to be the first of several projects where we're we're basically going to be opening up up the account console. Historically, it's been very, very account admin focused,
[13:23] uh and that leaves us in kind of a tricky situation where we know that a lot of people that are actually governance and platform members can't don't have access to the account console. So, we're going to change that.
[13:35] Uh so, here in this left-hand nav, you can see there's a new governance section. And I'll kind of click into every single one of these different views. Uh but just kind of focusing on the data one for a second. Uh here I can
[13:47] see the account level insights, but I can also click into a particular metastore. Uh and for Sorry? Oh, yeah, of course. Here. I know it's kind of pale in there.
[14:01] Is that a little bit better? Okay, cool. Hopefully. Okay, so here it's basically showing a drop-down view where you can see the account and then again can kind of click into a metastore. For the
[14:13] purposes of the demo though, I'll keep it account level just to make things easy. So, here in this view we can see there's a bunch of recommendations like what Michael was talking about before. And we have a bunch of widgets related to
[14:25] classification, how you're doing, and then how many quality quality checks are failing versus not. Um here I'm going to click in asset usage just cuz we were talking about that before. Clicking on this locations view, I can either view like the full
[14:38] list of tables based on how frequently they're used, or I can use this kind of interactive graph to see, okay, Fuzzer is the metastore that internally Databricks uses all the time. So, you
[14:50] can imagine there's quite a number of queries against data within that uh metastore. And then similarly, you can do the same breakdown for catalog, for schema, etc. The goal for this is for us to surface uh cases where you are
[15:04] missing tags, missing quality, and missing things that you need in order to ensure governance is enforced properly. There's also ability to view based on certified and deprecated candidates.
[15:16] All right. In the AI view, we kind of have a similar case where we have a bunch of widgets, we're showing you AI usage, spends. We talked a little bit about active users before, but I'll click into our model families one just to kind of show you the interaction
[15:28] pattern. So, basically here I can see grouping based on model family, how many requests are going through, how many tokens are associated, and how much spend is associated. Clearly, Databricks be using Llama a lot, so probably not too
[15:41] surprising. In the same view, I can switch between model families, users, workspaces all without leaving this page. In cost, we kind of have a similar case. We go through showing budgets, which
[15:55] ones are enforced, which ones maybe have been exhausted and need to be addressed, and we can show you actually when you're using compute and other resources that are untagged, which makes it a little bit trickier to
[16:07] actually attribute cost correctly. And then similarly, we have a spend analysis view in our cost explorer. So, here we can see revisiting that same topic of, "Hey, I want to see the spend
[16:20] in a given period per product line." I can see that here for all these different items related to jobs, apps, etc. And then also have the ability to filter by workspaces, tags,
[16:32] and basically whatever is most important for my analysis for that day. And then similarly for performance, so I'm super excited about this. This is a newer one. We basically are using this page to surface
[16:44] how you're doing in terms of performance across different warehouses with plans to include jobs and pipeline performance as well longer term. We also include a number of different optimization opportunities for, you know, table
[16:57] performance, health, and things like that. And then similarly, as promised, if I were to go here and click ask Genie, I can go ahead and start querying and asking additional follow-up questions directly here.
[17:10] I'll show you one that we had a conversation with last night that I sent to Michael just cuz I don't want to make you guys watch me make this load. But here we can see I asked for a breakdown of how I'm doing in terms of data
[17:23] quality across my entire account. And it spits out a whole analysis. Looks like something happened with the permission. Uh but basically I can see a summary view and fuzzer is the metastore we
[17:35] where we have the most usage and therefore the highest number of failing quality checks. All right, I'm going to pop back into the presentation.
[17:47] And then we'll talk through what's coming next. >> Cool. Thanks. Um Yeah, so what here in this section now we'll share a little bit about um where the governance hub is going in the in the future. Um so these are all
[17:59] additions we'll be making uh in the near term. So first we wanted to hit a little bit um on recommendations. So one thing I mentioned earlier is that right now uh actions are largely manual one by one,
[18:11] but we'll be adding the idea of agentic recommendations uh to apply uh action at scale. So Genie's able to actually draft actions uh for each individual finding. Um you'll be able to
[18:24] steer the fix to make sure the inputs are correct and review them and then be able to apply in bulk at scale. So instead of doing one by one, you can automate this across uh many different assets. Uh so while the GIF here is really nice, uh we actually do have like
[18:37] a little prototype of this. So uh I'll quickly show you all that instead.
[18:50] Okay. So here we are um in the governance hub. Uh so you can see we have our recommendations here. So I'll just choose one of them to start. So we'll choose missing uh comments on popular tables and we'll we'll click in.
[19:03] So now you have this view here where we have a native Genie view uh where we're able to have bulk uh action. So here on the right hand side of the screen, you'll see there is a uh an asset uh list that basically shows
[19:17] all the different findings you have across the assets uh in your account. In this case, there's 38 um and they're all selected. Um so, you're able to go ahead and actually control what's in scope to fix by uh selecting an individual set of
[19:30] them. Uh and for example, you could choose these three. Genie would be able to show you uh what actually is going on with the fix and what it's going to do. So, we can see in this case, uh there's three tables.
[19:42] We're going to add a one-line description to each table uh and Genie is able to actually draft the comments uh from the actual metadata. Um if you want more information on like an individual asset, you're able to to
[19:55] click and then Genie would go ahead and actually give you some stats and data about it. Um and also show you what the actual comment is. Um of course, you could also
[20:07] ask Genie to go ahead and change the comment. For example, maybe for this one we want to um make the comment on promotion codes
[20:19] um more uh table specific in this case. And then, Genie would go ahead um and make that that fix. Um additionally here, you'll have the
[20:32] ability to actually manage your recommendations through a life cycle. So, you'll be able to ignore uh recommendations if they're not relevant to you. As well as actually assign the recommendations to owners on the
[20:44] different assets. Um so, you can go ahead and choose that and it will send a notification to them in your destination of choice such as Teams, Slack, etc. And then, once you're happy with the recommendation, and if you actually want
[20:56] to fix it, you'll also be able to apply it here in bulk by uh clicking apply and then Genie will go ahead uh and apply these recommendations for you.
[21:10] Additionally, uh you'll have uh a single pane of glass. So, so far we talked about that there's a lot of different recommendations uh across um multiple domains. So, there's going to be one point of view to come in and see all the different recommendations you have across your entire account. Um
[21:24] and you can see all the different active findings underneath them as well as how they're trending over time, so you can see progress. There'll also be an actual history log across it, so you can see all the actions taken across your
[21:36] recommendations um by your team and what they resolved into. So, that's just like a little sneak peek of uh how we're thinking about making this a more full life cycle experience for recommendations um going forward.
[22:01] Yeah. So, I won't spend too much time here cuz we just we had the opportunity to look at it, but overall you could track the progress across your your recommendations and understand the current state and measure from a single pane of glass. So, then we'll we'll talk a little bit about
[22:14] also how Governance Hub is growing more broadly as well. Uh so, Governance Hub um currently is across multiple domains and we're going to be improving it across three separate vectors. So, Governance Hub will be uh getting
[22:27] better by becoming wider, deeper, and more open. So, by wider we'll be be including uh more domains uh such as uh right now we have available data, AI, and cost. We just added performance uh
[22:40] this week. And then we'll be adding uh additional domains like security uh by putting uh products such as like the security analysis tool uh into Governance Hub. We'll also be going deeper on our existing domains by improving the overall overall experience
[22:54] um by adding new features, uh for example, such as like access observability to the data domain. And then we'll also be expanding the Governance Hub by opening up uh access to it beyond admins in the account console to additional users who just
[23:07] have viewer permission. So, governance viewers. With that, I'll hand it over to Christian. >> Okay, cool. Um I'll zoom in on this, I promise. I know last time it was a little small. So, one thing I did want to highlight,
[23:20] and actually I'm going to go out of mode so I can actually zoom in on it. Hopefully that's a little bit better. Is uh we're actually introducing an access auditing view. I know this this one has been highly requested. So,
[23:32] I'm really excited about it. Uh basically you'll be able to actually see per a user group what permissions they have to assets across the platform that are governed by UC. Longer term, we will
[23:44] think about and kind of consider whether or not we want to include dashboards and some stuff that's traditionally in the workspace in here, but this will be a starting point. And we have resolved any issues where you can't see whether the
[23:57] permissions are inherited via a group. Two other ones I want to highlight, which are more representative of expansion for different verticals. So, luckily you guys were able to actually see the performance one live, uh which
[24:10] is kind of exciting. But, I did want to hit on this just quickly. Uh we are targeting introducing this late Q2. Sounds like we're a little bit faster, so hopefully sooner for you guys. Uh the main goal of this page is
[24:23] that we can actually help you diagnose issues with like your warehouses, uh figure out who's responsible for fixing them, and then also give you optimization opportunities. Uh there will be a lot of additional improvements in this surface as well as
[24:36] the other ones, though. Uh and then in addition to that, I'm actually curious how many of you are testing the security analysis tool right now? Yes? Okay, cool. Okay, nice. So, decent amount.
[24:49] Basically, we got a ton of really good feedback from that. Uh and we actually have the PM for this vertical in the room if you guys have follow-up questions. But basically we we got a lot of
[25:01] feedback on what you wanted to improve, what was working well, what wasn't, and we're actually taking that to heart and putting it as a productized version in governance hub as well. So you will have all of these same recommendations checks
[25:14] in addition to more like opportunities for root cause analysis in the future. >> Cool. And then I know this is a talk about governance hub
[25:26] but we also want to hit a little bit on system tables. Last year we had a a full session on system tables. This year we're going to have one slide. So I tried to condense this down to the most important announcements.
[25:38] So for system tables of course just to give everyone a brief refresher in case you're not aware, it's a programmatic way to observe data bricks. It's a basically a sequel it's a sequel it's a sequel interface for your observability data we provide. So in the last year
[25:52] we've made a bunch of progress. So we have four new platform features to share today. So for system tables we we have configurable retention which is currently in private preview and will be going to public preview
[26:04] later this year. It allows you to configure the retention on your system tables basically anywhere between 30 days and 10 years and also offers a free default retention of 395 days. So that will be
[26:16] broadly available soon. We've also been working on fine-grained access controls for system tables. So right now when you grant someone access to system tables they have access to all the rows
[26:28] and it creates an issue where you want to give someone access but it's very difficult cuz they're over permission potentially. So fine-grained access controls will limit the rows they can see to basically the rows they have permissions to for
[26:41] the individual assets. So as a concrete example would be if you're a workspace admin you'll only be you could only be able to see rows for workspace that uh you are an admin but not other rows.
[26:53] Additionally, we'll be introducing two other uh platform features that would just improve quality of life. Uh one of them is data latency SLOs. So, we'll be publishing uh completeness and freshness SLOs uh by uh system table
[27:06] um to provide certainty and just better understanding of how quickly data arrives. And we'll also be introducing low latency system tables uh later later uh this summer. Um with the first one being
[27:18] um the Spark declarative pipeline event logs. So, data in these tables will arrive uh will will arrive reliably in less than 5 minutes for near real-time monitoring use cases.
[27:30] And then just hit on a few of the tables. I won't go through all of them. But we also managed to release a large number of tables this year. Some of the big ones being AI Gateway where you're able to to see all the the central observability for that there as well.
[27:42] And tables for uh Zero Bus to understand um how that is working. And coming soon we have uh some of the most requested ones which will be our identity system tables um which are our principals, groups, and assignments. So, attribution will be
[27:55] easier. So, that's a a high-level summary. >> Oh, also just a plug. Uh we the govern tags one is in private preview right now. So, if anyone wants to test it before the public preview it,
[28:07] let me know. >> Cool. And with that just one other thing one other thing we have here is I you know, really love if everyone just pulls out your phone.
[28:19] I can all see uh Uh but yeah, no we really would appreciate if you pull out your phone, scan the QR code. There's just a quick Google Forms survey. Uh so, it's just an opportunity for you to provide us feedback on our road map
[28:32] so we can understand um what like what to prioritize most for you. Um so, we can deliver the best features uh next year when we all come back here. Um so, with that let's just take 30 seconds and just fill out the form.
[29:13] >> And then just a few other things um in terms of other sessions related to observability and governance um we have a few select ones here that we think would be highly relevant. Um if you're interested in this session uh some are very happened, some will be happening in the next few days. Um but I give these
[29:27] all a chance to go check them out. And if you didn't have the chance to catch the ones that happened yesterday or earlier today um go catch them on YouTube once the recordings are available. Uh it covers things like global UC
[29:39] data quality, uh what's new in terms of cost management and FinOps and then also Lake Flow jobs monitoring. So uh all very relevant sessions.
[29:52] And the last public service now announcement is to please complete your surveys in the Data and AI Summit app. Um it's really helpful feedback for us uh just for this overall event. So if you really like this session, give us five stars. If you hated it, give us one
[30:04] star. Uh all feedback's helpful. It goes into a long way in helping uh plan content.

Learn more about the Databricks Data and AI platform.

The information provided herein is for general informational purposes only and may not reflect the most current product capabilities or configurations.