LoginSign Up
← Back to all sparks
I

incident.io

INFRA · APIS
Velocity6.3

Incident management, on-call, and status page platform for engineering teams

incident.io ↗

incident.io Investigations doubles speed: 68% of incidents get accurate RCA in under 5 minutes

incident-managementai-rcaon-callalert-routingterraformnexus
◆Current state
incident.io's AI-powered Investigations feature — which auto-generates root cause hypotheses from the moment an incident is declared — just hit a significant performance milestone. Median time from incident open to an accurate channel message dropped from 6.7 to 3 minutes, and the share of incidents getting an accurate update within 5 minutes went from 8% to 68%. The supporting platform (bidirectional ticket sync, shardable alert rate limits, 24/7 schedule coverage) is being hardened around this flagship capability.
◆Where it's heading
Investigations is the product's primary differentiation and it's receiving sustained investment. The Terraform provider overhaul (resource by resource), alert source improvements, and schedule coverage policy fill out the enterprise-scale platform. The bidirectional ticket sync closes a key workflow gap: incidents and external ticketing systems now stay in sync without manual updates, which reduces the coordination overhead during active incidents.
◆Prediction
Investigations will likely expand to handle simultaneous incident streams and integrate directly with bidirectional ticket updates — turning the current 'hypothesis in channel' into an automated incident response loop that closes tickets as incidents resolve.

◆Recent moves

  1. 5d ago

    Investigations are 2x faster

    ⚡ SPARK

    Investigations drops its median time-to-accurate-channel-message from 6.7 to 3 minutes and raises the 5-minute accuracy rate from 8% to 68% — a 2x speed gain on the product's flagship AI feature, measured against real incident data.

    View source ↗
  2. 12d ago

    Bidirectional sync for incident tickets

    Incident tickets can now update the incident when the ticket changes, closing the feedback loop between external ticketing systems and incident state. Previously one-directional sync created drift; now changes in either system propagate.

    View source ↗
  3. 25d ago

    Easier alert source set-up

    Teams can now be assigned to an alert source regardless of what's in the payload — removing a configuration friction point that previously required payload-matching setup before routing could work.

    View source ↗
  4. 26d ago

    Shard alert source rate limits

    Alert source rate limits can now be sharded by a query expression extracting a value from incoming payloads — useful for multi-tenant or high-cardinality alert streams where a single rate limit would incorrectly throttle unrelated alert sources.

    View source ↗
  5. 1mo ago

    Reworking our Terraform provider

    The Terraform provider is being overhauled resource by resource — a sustained effort to make infra-as-code setup more maintainable. Not a single big-bang change, but a steady improvement track that reduces operational friction for enterprises managing incident.io at scale.

    View source ↗
  6. 1mo ago

    A few improvements to Status Pages

    Status pages get three additions: agent-driven updates, Pingdom uptime metrics integration, and self-serve language configuration. Each fills a gap that previously required workarounds or manual configuration.

    View source ↗