Failures, Retries, Cancellation & Heartbeats

When to use

Use for:

• •retry policy design
• •distinguishing retryable vs non-retryable failures
• •cancellation propagation (workflow → activity)
• •diagnosing “it keeps retrying forever”
• •long-running activities (heartbeats)

Required output

• •Recommended retry/timeout settings (what + why)
• •How to represent business failures vs transient failures
• •Cancellation-aware activity pattern (including heartbeat guidance if relevant)
• •Placement: what logic belongs in workflow vs activity

Guardrails

• •Do not hide failures; show explicit retry semantics.
• •Treat cancellation distinctly from failures/timeouts.