[EuiDataGrid] EuiOverlayMask needs a higher z-index when in fullscreen

[timroes]

The data grid fullscreen currently has a z-index of 8000, while the EuiOverlayMask has a z-index of only 6000. I think the EuiOverlayMask should have a higher z-index than pretty much everything else. Currently if you have actions in a data grid that will open a modal, this won't show when in fullscreen because it's hidden behind the fullscreen mask. I think it would make sense to still see modals even if the data grid is in fullscreen. I am not sure if there is any component where it would make sense, that it has a higher z-index than the EuiOverlayMask.

[cchaos]

We are starting to see this problem happen more frequently. Content in portals vs non-portals, elements with EuiOverlayMask as a parent vs sibling, z-index ordering not the same for all instances. (BTW, z-index layer maintenance is really hard when it's declared outside of the consuming app.)

I think most of it extends from the issue that the overlay mask and the content (modal) can be separated. If we add EuiModal as a child to EuiOverlayMask, then the overlay mask gains the same z-index level as the modal. And this would be the same for any component that uses EuiOverlayMask directly.

[timroes]

BTW, z-index layer maintenance is really hard when it's declared outside of the consuming app.

so much seconding this.

Could we potentially at least add a high level zIndex property to the EuiOverlayMask to give consumers an easier way of configuring it? (And maybe add a section in the docs, that you'll most likely need to adjust it to whatever you're putting inside it.)

[cchaos]

If you're using <EuiOverlayMask> directly (meaning it isn't added by an EUI component), you can always pass a custom className and supply your own z-index property. Since it is CSS, we try not to do inline styles (since they're hard to override) directly on EUI components and we can't have a prop and then matching styles for every single z-index value possible.

[cee-chen]

I believe this was fixed in #5054