incident.io vs Redocly

Investigations went generally available on 5 August, with Nexus posting a root-cause hypothesis and its evidence into the incident channel minutes after declaration. The releases since have been the operational surround: a 24/7 schedule coverage policy that flags gaps before a rotation goes unstaffed, more filtering in Insights, agent-written status page updates, and a resource-by-resource Terraform provider rework. This release lets an alert source extract a value from the incoming payload with a query expression and shard its rate limit on that value, so one noisy customer or cluster no longer consumes the whole source's budget.

Two threads run in parallel. Nexus started inside the incident channel doing diagnosis and has moved outward to the customer-facing artifact. The second is platform work for customers running at scale — coverage policies, escalation routing, and now rate limiting that understands the shape of the payload rather than treating a source as one undifferentiated stream. Sharded limits are what you build when a single alert source fronts many tenants, which says something about the size of the deployments driving the roadmap.

Rate limits that read a query expression out of the payload share machinery with routing and deduplication, so the same expression surfacing in alert routing rules is the likely next step. The changelog entry does not say.