Hi guys, We are getting more and more lags and downtime from the platform recently, sometimes it happens within the Dashboard page, sometimes it's Metrics page. Any idea of the root cause? Is there a way to increase the memory/compute allocated to our instance? We are just starting to scale the model, I'm starting to have concerns when we'll begin playing with billion of rows...

Hi Emmanuel, I’m sorry to hear about these problems. I’d be glad to help you troubleshoot them! Would you tell me a bit more about the kind of lags you’ve been experiencing? Do they affect some workspaces or dashboards in particular? When did they start occurring? I see that you’ve added a screenshot of a 500 error, which could have a different cause than the lagging you’ve described. What causes that one to pop up? Would you happen to have the

traceId

for that error? (it can be found under the error’s details, via the Browser Dev Tools, under the Console tab).

Hi @Francisco Antunes, thanks for the quick reply. We have been experiencing lags when we update metrics and save them (it says "saving" and load for a long time), until we are not willing to wait anymore, quit the page, load again and then get the "this page isn't working". Also, and maybe it's mostly when our devs are updating plugins at the same time, the loading of the dashboard takes forever. Feels like the caching mechanism is struggling or maybe we should tweak it? Because waiting more than 5s to load basic KPI cards feel a bit too much. Finally, for some workspaces, 2-3 dashboards can be saved anymore and we have no additional info on what's causing the issue. Only way to solve it was to duplicate the dashboard, start again and then it worked.

So I understand there seem to be 3 different issues here: 1. The long delays (that turn into 500 errors) when saving/updating metrics; 2. Loading issues with dashboards - which may occur only when plugins are being updated a. The plugin update explanation makes sense. I would recommend checking out our Cache Management options to ensure that the cache is cleared whenever such changes are made. 3. The issue with dashboards that can’t be saved on some workspaces. About #3, may I ask if these workspaces are part of the same tenant? Would it be possible that the dashboards in question are part of a Workspace Hierarchy, which causes them to become locked on the child workspaces, until new objects are created (via duplication)? As for the other two, please allow me to discuss this internally and I will come back with an update as soon as I’m able!

About #3, it's inside the same workspace where 2 out of 3 dashboards can't be saved anymore. They are all 3 at same level, no child-parent relationship. Another question @Francisco Antunes, don't know if it's better to post another message on the channel or just continue in this thread, but I don't find where I can access Usage Analytics (we purchased the feature). We need to start tracking ASAP users' behaviour on the platform. 🙏

Hi Emmanuel, You should have access to the usage workspace at the following URL: https://usageanalytics.cloud.gooddata.com/dashboards/#/workspace/cfua_walkway/dashboard/036df6f5-4b18-44e1-b6d5-885e635097e7 If you don't have access, we can help reach out through the proper channels to get you added there. Please let us know.

Hi @Joseph Heun, unfortunately the link does not work, it loads then results in HTTP ERROR 401.

I've requested access for you

Hello @Emmanuel Gautier, I have just enabled your access into Usage Analytics - you can log in now - please note that for the first time access, you will have to reset your password to set up new one.

@Marek Zelc amazing Marek, many thanks! How does it work if I want one of my coworker to use it as well? Can he access the url directly? (his name's Brent if you need it from the user base)

User access need to be provisioned by GD people at this point, so in case further person need access, you need to contact us (through your account owner, support or as you can see through Community 🙂 ). As for Brent, he is already enabled in the system as he was marked as initial user during environment provisioning.

That's perfect! Just a few questions on the definition of the KPIs please: • What's AFM? • Same for AFM executions? • AD Executions?

Sure thing: AFM (execution) - basically its any report computation (name it’s coming from API endpoint /executeAfm which is used for it. We are struggling here a bit how to name the metrics to be better self-explanatory). We are including this metrics as its “logged from backend” and do not suffer from different Ad-block and similar software that may affect telemetry collecting from other metrics, on the other hand, we do not have more granular info about such execution - the main idea is to use it as a benchmark for “how many users are blocking telemetry collection” and how reliable is granular data. AD Execution - execution done within Analytical Designer (Analyze tab in UI) Hope this helps.

Ok got it, all good for us, thanks!

Hi @Emmanuel Gautier, going back to the original issues here, could I ask for some additional info? • When did these issues start occurring? • Have you or your devs noticed any improvements in the last day or so? • Would you kindly share the URLs or IDs of the Workspaces, Dashboards, and Metrics on which the issues you described are occurring? (you can send them over via DMs, if you’d prefer) Thanks!

Hi @Francisco Antunes Problems did not occur since our last discussion so we are all good now, we also deleted the dashboard which were causing issues. Another quick question, does Usage Analytics refresh every day or every week? We are currently trying to check Usage but we can only see what was happening until 09/02/2024 (monday) Thank you very much for your help!

Hi Emmanuel, I’m happy to hear that the problems are resolved! Regarding the Usage Analytics, it should refresh daily, on the morning hours UTC. Depending on your timezone, the update for the 3rd of September could have brought in data from the previous day, from your perspective; In a few hours it should update again - if it does not, do let us know and we’ll look into it!

Hi @Emmanuel Gautier, thank you for reporting this. The delay is actually caused by mistake in data load configuration (activity data load is being finished after the overall ETL starts). I have corrected the setup now and starting tomorrow you will see data from previous day everyday after 6:30 AM UTC.

Thank you @Marek Zelc and @Francisco Antunes. I'm having some concerns about the Usage Analytics feature, could you please help me understand how can I get insights on what do users actually do or look at in the dashboard? All I see is basically just them loading it, but nothing more granular, meaning it's almost useless for us at this stage.

Hello Emmanuel, sorry for delay in answering, I probably missed notification about your message. As for granularity, the current granularity is set to day/workspace/user/object where object is either dashboard or report and these visits are recorded for loaded (displayed) objects. Deeper granularity is not planned at the moment. As for refreshes, those are set to happen daily, I’ll check whats happening

Hi @Marek Zelc, so if I understand it right, we can only check how many times a dashboard is loaded by a user, that's all? We can't see if they interact with filters like dates etc. ? That's not what was advertised when we agreed on the contract, as we mostly use one workspace for all of our users so we get no useful insights from it.

Yes, interactions with dashboards is the granularity level usage analytics currently provide. I have also noted your higher granularity requirements as feature request so we can consider adding it in the future. I’m not able to comment on your contract but I have highlighted the case to your account owner so you can discuss further.