Nexus aggregates real-time health data, activity events, scheduled job status, open requests, and token spend from every blueprint on the platform — without SSH, without tailing logs. One screen that tells you what is running, what failed, and what needs your attention right now.
The platform status panel gives you a single-screen health snapshot aggregated from every major subsystem. App counts show how many blueprints are registered and whether any have errored jobs. Kronos shows total scheduled jobs, how many are active, and error count. GHT shows tracked repos and last collection time. SRM shows open requests and today's volume. NLG and TKN surface newsletter output and total AI token spend. Metrics are sourced from the Sigma API when available and fall back to direct table queries — and the whole snapshot can be exported to CSV with one call.
The activity feed unifies the last 20 platform events across Kronos job runs, GHT collection runs, SRM request updates, SHL milestones, and freshly published Chronicle articles into a single chronological stream. Each entry shows the source app, a label, outcome status, and timestamp. Color-coded badges let you spot failures instantly without drilling into individual apps.
The Kronos section surfaces the health state of every scheduled job without leaving Nexus. Each job row shows its name, schedule, last run status, and a run history sparkline — five colored dots representing the last five run outcomes. Green is ok, red is error, amber is timeout, gray is idle. A job showing three or more red dots consecutive has already triggered a Telegram alert. Link to the full Kronos dashboard for Run Now and edit controls.
Any unacknowledged Herald alerts appear directly in Nexus. When a Kronos job hits three
consecutive failures, or another blueprint emits an error-level alert via
shared.herald.emit_alert(),
it surfaces here in real time. Source app, severity level, alert title, and timestamp are
all shown. Jump to Herald for the full acknowledge-and-triage workflow.
Nexus owns no tables. Platform metrics come from the Sigma API first and fall back to reading the underlying tables directly; the activity feed also pulls in SHL milestones and Chronicle articles. Every query is wrapped in a per-table try/except so a single unavailable service never breaks the whole dashboard.
kronos_jobs. Run dot history from kronos_runs.ght_repos. Last successful collection from ght_collection_runs.srm_requests. Excludes resolved/closed.nlg_newsletters. Unused pool from nlg_content_pool.tkn_turns across all AI sessions platform-wide.herald_alerts.Nexus owns zero tables. It assembles the full ops picture on demand from six live platform tables — no duplication, no staleness.
Apps, Kronos jobs, GHT repos, SRM requests, NLG content pool, and TKN spend — aggregated from six live tables into a single-screen status snapshot updated on every load.
The last 20 platform events from Kronos, GHT, SRM, SHL milestones, and Chronicle articles unified in one chronological stream. Color-coded badges surface failures instantly without drilling into individual apps.
Every scheduled job row shows a five-dot run history — green ok, red error, amber timeout, gray idle. Three consecutive red dots means an alert has already fired in Herald.
Unacknowledged Herald alerts appear directly in Nexus — source, level, title, and timestamp — so you spot active issues without navigating away from the ops dashboard.
A pass/fail tile for every registered app, built dynamically from the blueprint registry and refreshed by polling each /health endpoint. Newly added apps appear automatically; results are cached 60 seconds.
Download the entire platform status snapshot as a CSV — one row per metric — for spreadsheets, reporting, or feeding into other tooling.
Total AI token consumption and cumulative USD cost from TKN across all sessions platform-wide — visible at a glance on the status panel alongside Kronos and SRM metrics.
Every data source query is individually wrapped in try/except. A single unavailable table returns a zero and the rest of the dashboard continues unaffected — no cascading failures.
All 22 platform apps. Live health. Activity feed. Scheduled job dots. One screen. No SSH required.
Launch NexusRequires a MelTuc account. Create one free.