Two distinct TargetDown alerts are firing — job=pve (the revoked token, #250) and job=dns-servers — but both collapsed into one issue title because TargetDown has no namespace label, so the dns-servers case was never triaged.
Corrected evidence (an earlier version of this PR wrongly claimed the hosts were retired and the srv.renner.uno zone dead — that came from a broken resolver on the workstation, not the cluster): Prometheus's own scrape state shows context deadline exceeded for all three proxmox-dns-*.srv.renner.uno:12345 targets since 2026-07-13 — names resolve, the metrics endpoint doesn't answer. Root cause TBD by the fix agent once the issue exists.
Title key becomes namespace // job // \"cluster\" — verified live: [alert] TargetDown (pve) and [alert] TargetDown (dns-servers) are now distinct.
Migration after merge (I will do it): retitle #250 to [alert] TargetDown (pve); the next 15-min tick then files the dns-servers issue fresh for the hourly fix agent.
Two distinct TargetDown alerts are firing — job=pve (the revoked token, #250) and job=dns-servers — but both collapsed into one issue title because TargetDown has no namespace label, so the dns-servers case was never triaged.
**Corrected evidence** (an earlier version of this PR wrongly claimed the hosts were retired and the srv.renner.uno zone dead — that came from a broken resolver on the workstation, not the cluster): Prometheus's own scrape state shows `context deadline exceeded` for all three `proxmox-dns-*.srv.renner.uno:12345` targets since 2026-07-13 — names resolve, the metrics endpoint doesn't answer. Root cause TBD by the fix agent once the issue exists.
Title key becomes `namespace // job // \"cluster\"` — verified live: `[alert] TargetDown (pve)` and `[alert] TargetDown (dns-servers)` are now distinct.
Migration after merge (I will do it): retitle #250 to `[alert] TargetDown (pve)`; the next 15-min tick then files the dns-servers issue fresh for the hourly fix agent.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
https://claude.ai/code/session_01NV7bZzZLeoFUEdAeCSLEei
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Two distinct TargetDown alerts are firing — job=pve (the revoked token, #250) and job=dns-servers — but both collapsed into one issue title because TargetDown has no namespace label, so the dns-servers case was never triaged.
Corrected evidence (an earlier version of this PR wrongly claimed the hosts were retired and the srv.renner.uno zone dead — that came from a broken resolver on the workstation, not the cluster): Prometheus's own scrape state shows
context deadline exceededfor all threeproxmox-dns-*.srv.renner.uno:12345targets since 2026-07-13 — names resolve, the metrics endpoint doesn't answer. Root cause TBD by the fix agent once the issue exists.Title key becomes
namespace // job // \"cluster\"— verified live:[alert] TargetDown (pve)and[alert] TargetDown (dns-servers)are now distinct.Migration after merge (I will do it): retitle #250 to
[alert] TargetDown (pve); the next 15-min tick then files the dns-servers issue fresh for the hourly fix agent.🤖 Generated with Claude Code
https://claude.ai/code/session_01NV7bZzZLeoFUEdAeCSLEei
7d97cd6e60to74c2a2a75c