[vimtor]

Removes the SST-managed Honeycomb triggers, the Console /honeycomb/webhook route, and the LogProcessor tail worker.

The triggers query event_type = completions, which has had no data in Honeycomb for 90+ days since the log processor stopped exporting. Alerting now lives in the hand-made triggers on the inference dataset. Confirmed with Daniel.

LogProcessor lost its last destination when the S3 lake was retired (#52514) and only console.logs to a worker nobody tails. The Console no longer emits _metric: logs since /zen/* proxies to inference.

• delete infra/monitoring.ts, infra/secret.ts, webhook route
• drop DISCORD_INCIDENT_WEBHOOK_URL, HONEYCOMB_API_KEY, HoneycombWebhookSecret links
• delete LogProcessor worker, Console tailConsumers, and packages/console/function

The honeycomb and random providers and HONEYCOMB_API_KEY in deploy.yml stay so the next production deploy can delete the resources. Remove them in a follow-up.

[opencode-agent]

I found no problems with this change. The console function and console app type-check cleanly. Nothing in the code still uses the removed resources or the /honeycomb/webhook route. I ran the old and new log processor tail handlers on the same fake event and they printed identical output, with later metrics still overriding earlier ones. Keeping the honeycomb and random providers and HONEYCOMB_API_KEY in deploy.yml means the next production deploy can still delete the triggers and the RandomPassword resource. I couldn't run a real deploy here because there are no credentials.

Public API: unchanged. This only touches infra, the internal Console webhook route and the log processor worker.