Some solution needed for when topology needs to display too large set of data

Question

Some solution needed for when topology needs to display too large set of data

abonas opened this issue 9 years ago · 6 comments

abonas commented 9 years ago

perhaps some message of "too much data to display" ?

Answer 1 · 2016-03-28T09:37:06.000Z

@serenamarie125, any suggestion how this case should be handled? thank you

Answer 2 · 2016-03-29T12:11:01.000Z

Similar request in Cockpit: cockpit-project/cockpit#4041 I think @andreasn will be working on this.

Answer 3 · 2016-03-29T12:17:55.000Z

Yeah, this looks interesting.
One of the main issues in cockpit-project/cockpit#4041 is that it's hard to tell what pod/cluster/service/etc is what. So in that case I was kind of thinking of adding more data, like a name to each object, and that wouldn't really help in this case, quite the opposite. :)

Answer 4 · 2016-03-29T12:19:02.000Z

I left a comment here with some ideas based on enhancements I added on miq side to topology that can solve several of the issues brought up there in regards to search and names of objects:
cockpit-project/cockpit#4041 (comment)

Answer 5 · 2016-03-29T16:42:08.000Z

@abonas I would suggest that if you detect that issue, you set filters so that only a subset of entities are shown by default, and then show a modal stating something like

Based on the size of your environment, filters have been set to show only nodes by default.

and have the modal only have a close button

(i'm not sure that nodes are the appropriate object types to show by default). Would that make sense?

Answer 6 · 2016-03-30T09:28:20.000Z

@abonas I would suggest that if you detect that issue, you set filters so that only a subset of entities are shown by default, and then show a modal stating something like

Based on the size of your environment, filters have been set to show only nodes by default.

and have the modal only have a close button

(i'm not sure that nodes are the appropriate object types to show by default). Would that make sense?

I think the technical obstacle here is to determine "how much is too much" - how many entities look good on graph and go from there. from functional perspective your suggestion is great. regarding what to display, I think it depends on what the topology will be focused for (future "per entity topology" might not be necessarily be focused/include many nodes)