Founders building AI products: how do you monitor critical events?

AI apps generate tons of events: model errors, API limits, automation failures, background jobs, and system alerts.

But sending everything as notifications quickly becomes noise.

How do you currently monitor important events in your AI stack?

• Observability tools?

• Slack alerts?

• Custom dashboards?

We’re experimenting with a rule-based alert system in NotiLens, but I'm curious what others in the Product Hunt community use.