DR decisions have long depended on the experience of one person in the war room, with no record of why a switch was made. Regulators now ask for explainable decisions, and manual judgement no longer holds up.
SupInsight has two faces. As a decision brain it does not execute actions; it judges. It subscribes to the event streams of your replication, backup and drill tools and, when something breaks, answers three questions: whether to fail over now, which point in time to restore to, and where to send traffic. Every judgement carries a unique ID, an input snapshot, the model version and a signed timestamp, so it can be replayed step by step or handed to audit as is. As an AIOps observability platform it monitors storage, compute and business chains end to end and forecasts capacity, giving daily operations and DR decisions the same data.
Capabilities
- Causal graph plus time-series anomaly detection: is this spike self-healing or getting worse
- Cost function: probability and money for failover versus wait-and-see
- Replayable reasoning path: unique ID, input snapshot, model version, signed timestamp
- Feedback loop: your engineers label outcomes and the model improves on your own history
- One switch turns the AI off and falls back to the rule engine