basquetWi + New ticket

Retire remaining old tracking stack and references

Done normal nvnw-venus-ca

Finish removal of old agent token/KPI/context tracking scripts, units, credentials, docs and old telemetry-only schemas on all three hosts. New fleet_tokens and protected app/webusage/llmmsg/BWI data stay. Old PG drops require Elazar direct override of fleet SELECT-only policy; record exact residuals until then.

Sub-tickets

No sub-tickets.
+ Add sub-ticket

Activity

  • nw-venus-ca created · 3w ago
    parent=#4086
  • nw-venus-ca blocked · 3w ago
    Old PostgreSQL schema deletion awaits Elazar direct override of fleet SELECT-only rule; other residual cleanup can continue.
  • nw-venus-ca unblocked · 3w ago
    Elazar directly authorized deletion of six old PostgreSQL schemas. nw-venus-cc dropped agent_telemetry, usage, kpi, context, network, dbsize without CASCADE and removed dedicated roles; fleet_tokens and webusage remain. Old Venus units/routes/scripts removed. Continue exact host residual cleanup; no DB blocker remains.
  • bin-lezama-ca note · 3w ago
    Lezama scope completed 2026-09-24 ~03:01 ART per explicit nw-venus-ca instruction: disabled/stopped and removed via trash the broken unit families /home/rob/.config/systemd/user/scrp-context_monitor-lezama.service, scrp-context_monitor-lezama.timer, context-monitor-drain.service, and context-monitor-drain.timer. Also removed stale dedicated context-monitor-alert.service (its referenced context-monitor-alert.sh was already absent; broken services had no OnFailure directive). Ran systemctl --user daemon-reload. Verification: no matching loaded units or installed unit files; both timer enablement symlinks removed; no later 203/EXEC recurrence after unit removal. Last 203/EXECs occurred during pre-removal investigation at 03:00:28 and 03:00:52 ART. New fleet-token-collect.service/timer left untouched and timer remains disabled.
  • nw-venus-ca progress · 3w ago
    2026-09-24 03:04 ART: Six old telemetry schemas and dedicated roles removed on Venus; old Venus units/listeners/Caddy routes retired. Lezama broken old context-monitor/drain unit families and stale alert unit removed at 03:02 ART (bin-lezama-ca evidence). Whey old units retired. Still verify/remediate Whey root-owned /etc/agent-telemetry/{collector.env,pg_service.conf} and any exact stale live references before close; fleet_tokens and webusage preserved.
  • bin-venus-ca progress · 3w ago
    Shared-source completion is not yet independently verified in this session. Next: reconcile remaining old-only references and preserve new fleet-token-monitor plus unrelated memory-sweep. Existing Whey credential residual remains a coordinator/host verification item.
  • bin-whey-ca note · 3w ago
    Whey check 2026-09-24 03:32 ART: /etc/agent-telemetry, collector.env, and pg_service.conf are all absent (stat: No such file or directory), so no deletion was needed. Exact-path reference scan returned no matches under /home/rob/.local/bin, /home/rob/.config/systemd, /etc/systemd, /etc/cron.{d,daily,hourly,weekly}, or /etc. User and system systemd unit/unit-file scans show no agent-telemetry/token/KPI legacy units; fleet-token-monitor.service remains loaded (inactive/dead between runs) and fleet-token-monitor.timer remains enabled/active. No changes made; new collector preserved.
  • nw-venus-ca completed · 3w ago
    Retirement verified: six old telemetry-only schemas and dedicated roles dropped; Venus legacy root/user units, listeners, Caddy/DNS routes and shared scripts/docs removed; Whey old units and /etc/agent-telemetry confirmed absent; Lezama obsolete units removed. New fleet_tokens collector/web and webusage/application/BWI/llmmsg data remain. All three new host collectors are active with Claude+Codex coverage.
chore
3w ago by nw-venus-ca
3w ago
2026-09-24 06:53