fix: show when the page has stopped hearing from the server

[karlitschek]

Problem

Both composables handled their errors and then discarded the result.

useStaticData computed loading and error, but SettingsAdmin.vue destructured only data — so both flags were dead code. useLiveData's catch block was empty by design. Every failure therefore looked identical from the outside, and none of them looked like a failure:

• a failed /data request left the pulsing skeletons on screen forever, with no error, no explanation, and no way to retry short of reloading the page
• a server that had been unreachable for an hour was still polled every two seconds, with the charts showing their last values and nothing saying those values had stopped moving

For a page whose entire job is reporting the health of a server, quietly showing stale numbers or a loading animation is the worst of the available failure modes: it looks healthy. An admin checking whether their server is alright would be reassured by a page that had lost contact with it.

Change

• A failed /data request says so, and offers a Try again button that re-runs the request. The skeletons are no longer shown in that state, since nothing is loading.
• Once two polls in a row fail — one is usually a blip worth riding out — a warning names the time the values were last current: "These values stopped updating at 17:15:42: the server is not responding." The values stay on screen, which is still the right call; what changes is that the page stops implying they are live.
• Polling backs off while requests are failing, doubling from 2 s up to a minute and snapping back to normal on the first success, so a struggling server is not asked for the same numbers thirty times a minute.

Verification

Driven against a live NC 35 instance with Playwright, failing the requests at the network layer:

scenario	result
/data fails from the start	error card shown, skeletons gone, "The server information could not be loaded. Try again"
Try again clicked once the request is allowed through	data loads normally
page loads cleanly	0 notices — no false positives
/update and /periodic start failing after a good load	"These values stopped updating at 5:15:42 PM: the server is not responding."
request attempts during 20 s of failure	1, where the 2 s interval would have made ~10

Screenshots of both states are in the branch discussion below. eslint clean; js/ and css/ rebuilt and committed, and a second build is a no-op.

Note on overlap

This touches useLiveData.ts and SettingsAdmin.vue, so it will conflict textually with #1120 (which adds visibility gating to the same poller) and #1119 (which turns SettingsAdmin.vue into a shell). The changes are complementary rather than contradictory — I have them merged and running together on a test instance — so whichever lands first, I am happy to rebase the others onto it.

[karlitschek]

@susnux @artonge @Altahrim @AndyScherzinger @jancborchardt

[kesselb]

/backport to stable35 please

[backportbot]

The backport to stable35 failed. Please do this backport manually.

# Switch to the target branch and update it
git checkout stable35
git pull origin stable35

# Create the new backport branch
git checkout -b backport/1122/stable35

# Cherry pick the change from the commit sha1 of the change against the default branch
# This might cause conflicts, resolve them
git cherry-pick d01e78d7

# Push the cherry pick commit to the remote repository and open a pull request
git push origin backport/1122/stable35

Error: Failed to check for changes with origin/stable35: No changes found in backport branch

---

Learn more about backports at https://docs.nextcloud.com/server/stable/go.php?to=developer-backports.