Why does our FR10 availability tile turn green after a failed request?
LinBaker0444 · 17 Sept 2025, 09:12 UTC
19 replies
Does the tile change to green at the failure, or stay green from an earlier successful reading?
22pointsChanges from unavailable to green in our saved replay. Not just an old colour hanging around. The raw failed reply is saved with that run.
6pointsSend the maintainer that reply with the installed SDK version. He needs the documented response format, not a guess at which field looks like status.
7pointsWe had a monitor doing this with the success of receiving a reply. It had heard something, therefore all was apparently well. Ask him to test the failure cases directly, including missing or malformed responses.
13pointsIs this only a display, Lin? Or does another part of your bench software use that available value?
13pointsOnly the display as far as our maintainer's trace goes. He found the whole response being treated as true or false, which explains why a failed request with information in it could light the tile.
14pointsI'd blank the availability value when the request fails. We use a separate communication indication; otherwise people spend time asking whether unavailable means the controller answered or nobody could reach it
8pointsCould it keep the last good value with its age? That would help me describe what happened before the failure, though I'd want it clearly marked old.
15pointsGabriel, I'd like that in the detail view. Julia's blank current value would be less confusing on the main tile. Our team mostly glances at the screen from the other bench.
6pointsPlease don't leave the old green background behind a tiny stale label, because from the other bench nobody is reading your tiny label.
6pointsWho gets the failure detail when a technician opens the tile?
13pointsAnyone using the bench can open it. Alex, we're adding the request outcome and last successful reading time there. Oscar, agreed on colour; our mock-up now has a distinct no-current-reading state.
2pointsKeep the raw diagnostic detail available without making it the main label. Good for support, fairly miserable as an operator message.
16pointsReplays now cover success, reported failure, a missing response and malformed content. Maintainer checks the documented success result before taking the status field. Failed cases don't get a current availability value. Thanks Sofia, asking whether it changed colour made me look at the replay properly.
9pointsAnd recovery? The first valid response should put it back without somebody reopening the dashboard
18pointsI'm happy with the history in the detail view. Better than trying to make one green square tell two different stories.
13pointsDid the installed version behave like the replay?
19pointsYes, Jonas. Maintainer's bench check covered loss and return of valid readings. Julia, it recovered without reopening anything. The corrected display is installed; we also removed the old true-or-false helper so it doesn't get borrowed for another tile.
3pointsUseful ending. I'll ask our maintainer about that helper too.
8pointsDiscussion closed
This discussion is closed to new replies after six months without activity. Last activity: .