Delivery retries: Delivery and operations

Specifies queues, providers, retries, deduplication, tracking, observability, fallback, and failure handling for delivery retries. It fixes adoption boundaries and observable failure evidence while keeping environment-specific limits configurable.

Package status: reference context ready for human review. The contract and test scenarios are complete, but no claim is made that an adopting implementation has passed them.

Decision

For Delivery retries, adopt a durable notification intent identified independently of provider attempts. Operate the capability with named signals, bounded capacity, owned alerts, a diagnostic runbook, and tested recovery evidence. This package is a reference contract: the adopting team must choose numeric limits, providers, jurisdictions, and operational owners from evidence in its own environment.

Scope

• The Delivery retries actors, inputs, outputs, states, policy or schema versions, and externally visible outcomes described by this operations block.
• Primary and alternate interfaces, background work, caches, integrations, support paths, and evidence that can exercise the same boundary.
• Adoption-specific configuration, rollout, recovery, and verification responsibilities needed to apply the reference safely.

Outside this block

• Choosing a universal vendor, framework, jurisdiction, numeric threshold, retention period, or service-level objective for every adopter.
• Claiming that packaged scenarios have run against a downstream implementation or that this reference grants legal, security, or accessibility certification.

Contract

• Create one notification intent with recipient, purpose, template version, locale, schedule, sensitivity, and deduplication identity.
• Apply verified recipient state, channel preferences, quiet hours, and legal or product suppressions immediately before dispatch.
• Track queued, attempted, accepted, delivered where observable, failed, suppressed, and cancelled as distinct states.
• Bound retries per provider and preserve the same intent identity so a fallback channel cannot create duplicate user actions.
• Define normal, degraded, exhausted, and unavailable states with decision metrics and explicit operator actions.
• Preserve accepted-work identity during failover, replay, rollback, and repair so recovery cannot duplicate or silently lose effects.
• Separate one durable notification intent from provider attempts and classify validation, permanent rejection, transient failure, throttling, timeout, and ambiguous acceptance.
• Retry only safe classes with exponential backoff, jitter, attempt and age bounds, stable deduplication, provider reconciliation, and an inspectable terminal state.

Implementation guidance

• Practice the runbook in a representative environment and record recovery time, data loss boundary, and unresolved assumptions.
• Model Delivery retries inputs, outputs, actors, states, invariants, side effects, and evidence before selecting framework or vendor details.
• Store the applicable Delivery retries contract or policy version with material state so migrations, replay, and support decisions remain attributable.
• Introduce the path behind controlled rollout, compare expected and observed outcomes, and keep a tested recovery route until adoption evidence is complete.

Failure handling and safeguards

• Permanent address or policy failures suppress further attempts; transient failures use bounded retry and then an operator-visible terminal state.
• Invalid or contradictory Delivery retries state is rejected or quarantined; the implementation never guesses a value that broadens authority or duplicates an effect.
• Retries are bounded and use durable identity; after exhaustion, work reaches an inspectable terminal state with an accountable owner.

Verification and operations

• Inject saturation, dependency outage, stale configuration, and telemetry loss and confirm the declared degraded behavior.
• Monitor Delivery retries success, denial, validation failure, dependency failure, retry exhaustion, and recovery by contract version without sensitive payload dimensions.
• Re-run paired positive and negative fixtures after policy, schema, dependency, migration, or boundary changes and preserve the resulting evidence.

Adoption assumptions

• Names and values in the example are a concrete fixture, not universal defaults; adopters replace them through documented evidence and ownership.
• The adopting system can provide authenticated identity, durable operation or record versions, bounded telemetry, and a controlled path for change.

The executable-looking examples in this package are fixtures and acceptance contracts. Run the collection validator to check structure and metadata, then translate and execute the scenarios in the target repository before recording implementation evidence.