Workflows that don’t die.Scale-out workers. Guaranteed state. Repeatable outcomes.
Immortal coordinates workflows, activities, retries, and timers across a hub-and-spoke network — with security and observability built in.
A hub-and-spoke worker network
Immortal routes tasks through a central server while workers stay lightweight. Edges transmit data continuously — perfect for multi-tenant queues and bursty workloads.
- Single upstream stream per queue
- Backpressure-friendly fanout
- Worker registration + discovery
Secure workflows end-to-end
Validate identity, authorize execution, and protect payloads in transit. Bring your own IdP, keys, and policies.
- mTLS between nodes (optional)
- JWT / OIDC verification
- Request signing / key rotation friendly
Explainable policy decisions
Whether you use RBAC or relationship-based permissions, Immortal can evaluate access before execution and explain exactly why a decision was made.
- RBAC / ReBAC / ABAC compatible
- Decision logging (allow/deny + reason)
- Tenant-safe by design
Retries you can trust
Configure retry policies per activity. Record every attempt. Support idempotency keys to make retries safe.
- Exponential backoff + jitter
- Per-step retry policies
- Deterministic history for debugging
Timers & sleeps without holding compute
Pause a workflow, persist state, and resume later — great for follow-ups, delayed tasks, and long-running business processes.
- Durable sleep/resume
- Optional: kill compute while sleeping
- Accurate sequencing in UI history
Worker leases & heartbeats
Workers claim tasks with leases and send heartbeats. If a worker stalls or disappears, work is reassigned automatically.
- Fast dead-worker detection
- Avoid duplicate work with leases
- Built for autoscaling
A timeline you can replay
Every important event becomes a first-class record: scheduled, started, attempts, sleeps, resumes, outputs.
- Deterministic history
- Human-friendly UI ordering
- Audit and incident-friendly
Stream logs now, replay later
Tail workflow logs in real time and rewind using history-backed context. Great for audits, debugging, and customer support.
- Server-aggregated streams
- Workflow-scoped log views
- Replay highlighting from history