AI OPS
LiveCloud and AI spend, the five support automations, and the operational state of the pipeline behind them.
All five modules now run in Spark. Cost, Services and Provider health read the AI-OPS cost database directly; Automations reaches support-sync through a narrow, fixed set of forwarded requests. aiops.xops.site is no longer needed to operate any of this.
Daily spend across AWS, Cloudflare, Sentry, Anthropic and OpenAI, with per-provider units, a clickable breakdown on every dimension, and CSV export.
The five support automations, each with its scope, run log, dry run and documentation. Two run partially; three are built but switched off.
What the scheduled pipeline last did, per service, and when.
Whether each cost provider is still returning data, so a silent collector is visible rather than looking like a quiet week.
How the platform is put together: the workers, the databases, the schedules, and what each automation actually touches.