# Incidentary > Shared incident tracing for distributed systems. The causal chain across your services, assembled in seconds, visible to everyone in the war room. Incidentary captures the distributed execution trace in the 60 seconds before an alert fires and delivers it as a single shared artifact. When something breaks, every engineer on the team looks at the same deterministic causal chain instead of five different dashboards with five different theories. ## What it is Incidentary is a causal trace layer that sits between observability infrastructure (Datadog, Grafana, Honeycomb) and incident management tools (PagerDuty, incident.io, Rootly). It does not replace either category. It fills the gap between them: when the alert fires and the team assembles, Incidentary provides the shared causal picture that no coordination tool or dashboard produces. The trace is built from actual parent-child propagation headers (parent_ce_id) through every service boundary — HTTP, gRPC, queues. This is deterministic causal assembly, not probabilistic correlation or statistical inference. Two engineers looking at the same trace link see the same determined sequence of events. ## Who it is for Engineering teams at companies running 3-15 distributed services with 10-50 engineers. Typically Series A/B startups and growth-stage teams that have recently outgrown a monolith. Primary users: SREs, on-call engineers, DevOps leads. ## Key capabilities - Pre-alert capture: 60-second ring buffer captures the causal chain before the alert fires, not after - Deterministic causal assembly: real parent_ce_id propagation, not probabilistic correlation - Ghost service detection: single SDK install reveals uninstrumented dependencies in your topology - Shared trace links: no-auth URLs — paste in Slack and the whole channel sees the same timeline - Coverage scorecard: continuously surfaces instrumentation gaps before incidents expose them - OTLP-compatible: existing OpenTelemetry traces can feed in ## SDKs SDKs for Node.js, Python, Go, and .NET. All Apache 2.0 licensed. Auto-instrumentation with zero code changes — one middleware install per service. ## Pricing Usage-based pricing per causal event (CE), not per seat. See /pricing.md for machine-readable pricing data. - Free: $0/month, 200K CEs, 14-day retention, 5 members - Pro: $59/month, 1M CEs, 30-day retention, unlimited members - Team: $149/month, 5M CEs, 90-day retention, unlimited members - Enterprise: custom pricing, custom retention, SSO/SAML, dedicated support 1 CE = 1 captured operation. A typical 3-service request with 2 downstream calls produces ~6 CEs. ## Key pages - [Homepage](https://incidentary.com/): Product overview and quickstart CTA - [Quickstart](https://incidentary.com/docs/quickstart): 5 minutes to first trace - [How it works](https://incidentary.com/docs/how-it-works): Technical architecture - [Why Incidentary](https://incidentary.com/docs/why-incidentary): Problem statement and positioning - [Pricing](https://incidentary.com/pricing): Plans and CE-based metering - [Ghost services](https://incidentary.com/docs/concepts/ghost-services): Uninstrumented dependency detection - [Pre-alert capture](https://incidentary.com/docs/concepts/pre-arm): Ring buffer and pre-alert trace assembly - [Causal events](https://incidentary.com/docs/concepts/causal-events): The unit of capture - [Blog](https://incidentary.com/blog): Technical writing on incidents and distributed tracing ## SDK documentation - [Node.js SDK](https://incidentary.com/docs/sdk/node) - [Python SDK](https://incidentary.com/docs/sdk/python) - [Go SDK](https://incidentary.com/docs/sdk/go) ## Integrations - [Slack](https://incidentary.com/docs/integrations/slack): Trace link unfurls and incident notifications - [PagerDuty](https://incidentary.com/docs/integrations/pagerduty): Alert-to-trace correlation