Hey Tom, I'm having an issue where the following 2 services are exited. Even after removal of the Metrics & Dashboard app These won't remove.
5 replies
tomspark wrote:
Those two are harmless leftovers — they're tiny one-shot setup helpers (a Prometheus and a Grafana 'prep' container) that run once, exit, and normally get cleaned up. When you removed the Metrics app they got left behind in the 'exited' state because they'd already stopped. They're doing nothing, but let's clear them out: docker rm -f sb-prometheus-init sb-grafana-init sb-prometheus sb-grafana sb-node-exporter sb-cadvisor That force-removes any metrics containers still hanging around (it's fine if it says 'no such container' for some of them — that just means they were already gone). After that they'll be off your list for good. Quick one so I can tighten this up: did you remove Metrics from the dashboard's app page, or with 'sparkbox disable metrics' in the terminal? Want to make sure the cleanup catches these automatically next time.
Big_Totoro wrote:
Tom, I removed through the dashboard app page. Note that even after the initial install these 2 containers stayed and showed exited and were counted as services that needed attention on the main page.
tomspark wrote:
Yeah — those two are the one-time setup helpers (sb-prometheus-init and sb-grafana-init). They run once when the module installs, then sit there showing 'exited', and you're right that the dashboard is wrongly counting them as services that need attention. Two real gaps on our side: they should clean themselves up after they finish, and the health check shouldn't flag a setup helper that already did its job. I'm fixing both. To clear them right now: docker rm -f sb-prometheus-init sb-grafana-init They won't come back. Thanks for the detail, that's genuinely helpful.
Big_Totoro wrote:
Hey Tom, Just did a removal and reinstallation to test if the helpers get cleaned up. They don't. Every update or recreation of the parent containers brings back those init helpers and they stay and are still counted as items that need attention. Edit: I missed that you said you are working on a fix. My apologies. I'm on 1.6.312
tomspark wrote:
No apology needed — and that's exactly the detail that matters: you've confirmed they come back on every recreate AND keep getting counted as needing attention, even on 1.6.312. That's the real bug, and it's logged with your repro so the fix covers it properly (both the lingering helpers and the dashboard wrongly flagging a setup container that already did its job). You don't need to keep removing/reinstalling to test — I've got what I need. For a clean board in the meantime, docker rm -f sb-prometheus-init sb-grafana-init clears them until the fix ships. Thanks for being thorough with this.