Hello :wave: Have you come across browser crashes ...
# gooddata-ui
t
Hello 👋 Have you come across browser crashes when viewing stacked bar graphs? It happens to us with “busier” graphs. If you move your mouse over the bars, or change the size of the screen, browser crashes. Interestingly, this seems to happen only in V8 based browsers. E.g. Firefox seems to handle this fine. It appears to us as a memory leak. Looking at the system monitor, the RAM used by the process quickly blows up before the page crashes. When I run chrome with debug logging, we get a message saying
V8 process OOM (Oilpan: Large allocation. Ran out of reservation)
.
m
Hi Tomas, I just wanted to let you know that I am going to have a look into this and get back to you asap
🙌 1
One quick question - Do you see this on the dashboard once the report is placed there? Or you see this in Analytical Designer when making reports?
Can you please also tell me which version of Chrome you’re using and if you have tried to check for any updates? 🙂
t
Ah my bad, I forgot to add that this happens only when we use UI SDK (React) to embed the dashboard within an application. In the native UI it works fine. We have tried to remove all our custom code to have the most minimal setup of the Dashboard component to ensure we are not causing the problem with our code. Chrome is latest (128) running on MacOS.
👀 1
and to narrow it down further, the same issue happens when embedding the chart using the
InsightView
component only
👍 1
if it helps, we could share the insight definition and its data as well
m
Hi Tomas, can I ask if you’re using the latest version of the GoodData UI SDK 10.8?
t
yes, we’re using the latest one
m
Hi Tomas, I went ahead and tried to replicate this behaviour, but I was not able to reproduce it on my side using our React Quick Start App. I went ahead and created a very “busy” graph and used UI SDK (React) to embed my dashboard using UI SDK 10.8 via in Chrome and I was not able to replicate the issue on my side and when looking at the Bar graph and the at the RAM used by the process there was no spike in activity at all on my side. So in this case, I would like to kindly ask if you (if it will be possible and not too much trouble) to try and use our React Quick Start App too and see if you’re able to replicate this issue there in that environment?
t
sure, I can do that. is there a way to take the graph definition a data from our production environment and feed them directly to the InsightView or Dashboard component?
m
t
that’s actually what I’d like to avoid. I don’t want it to be connecting to our prod GD Cloud, but just statically feed the the data and definition to that graph so that I can share it with you. is there a way?
I mean, I could mock the network requests, but if there’s a more straightforward way, it’d be nice
m
In this case, would it be possible to DM me and send me your .json package and with this I will escalate this case internally who will then try and replicate your setup on side?
Hi Tomas, just wondering if there is any update on this issue? As I mentioned previously, if you would like our team to investigate this issue further, would it be possible to DM me and send me your .json package and with this I will escalate this case internally
t
Hi Michael! Thanks for checking back. My plan was to try and built something reproducible for you to see, but I didn't have time to finish it off yet. But we're still experiencing crashes and yesterday it happened on another type of graph as well. I'll DM you later today with more info. Thank you!!
m
Hi @Tomáš Gajdoš I just wanted to let you know that I am heading off for vacation shortly 🌴🥳 and I just wanted to let you know that my colleague @Julius Kos will be taking over this case now - So please DM him directly with the mentioned additional information and my colleague will review it and if needed, escalate the issue internally for you 🙂
❤️ 1
t
Hi Michael! Just for the info, I found the cause in the end. We use Radix UI in our app and if the Dashboard/Insight was rendered within it’s Theme provider component, it’d experience performance issues/memory leaks/crashes. I think it has something to do with the fact that the provider defines large number of CSS variables on a class that is assigned to the wrapper along with some more complex CSS selectors and media queries. With the large number of DOM nodes found in these Insights, it ends up crashing (in Chromium browsers only). This is some low level stuff, most likely a Chromium CSS engine bug and I couldn’t figure out precisely what the cause is, but we can work around it. So all good. Thanks for your help.