Secrets Management — System Design Academy

Complete System Design Tutorial

The design separates the latency-sensitive online path, the correctness-critical durable transition, and asynchronous projection or repair work. Each chapter below explains one decision shown in the diagrams.

1. Requirements and scope

Functional requirements

Execute the primary secrets management workflow and return an observable result.
Expose the concrete operations shown below with stable retry identities.
Allow operators to inspect, disable, repair, or replay ambiguous state.
Keep derived analytics and notifications off the correctness-critical path.

Non-functional requirements

Design for Millions secrets · 100K reads/s · HSM-backed root keys.
Protect the named business invariant under concurrency and partial failure.
Scale by an explicit aggregate/partition key without one global mutex.
Degrade intentionally when cache, queue, store, or regional ownership is unavailable.

Primary invariant: each accepted secrets management operation has one stable result and retries do not repeat its business effect.

↑ top

2. Capacity assumptions

State assumptions before choosing technology. The supplied workload is Millions secrets · 100K reads/s · HSM-backed root keys. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	Millions secrets · 100K reads/s · HSM-backed root keys	Separate online compute, durable state, and asynchronous consumers.
Peak factor	5–10× average	Admission control, autoscaling headroom, and hot-key protection.
Stored state	write rate × retention × record bytes	Partition count, indexes, compaction, backup, and archival tier.
Network	peak operations × payload size	Batching, compression, edge placement, and egress budget.
Availability	99.9–99.99%	Multi-zone replicas, tested failover, and dependency error budgets.
Recovery	RPO/RTO by invariant	Ack boundary, replay retention, reconciliation, and operator runbooks.

Do not cargo-cult the numbers: show the arithmetic, name the uncertain variable, and say which production metric will replace the interview estimate.

↑ top

3. API and data model

POST /v1/secrets/{path}/versions

GET /v1/secrets/{path}authorized lease

POST /v1/credentials/{role}dynamic secret

POST /v1/leases/{id}/revoke

Authoritative records and indexes

secret(path, version, ciphertext, key_ref, state)

policy(principal, path_pattern, capabilities, version)

lease(id, principal, secret_version, expires_at)

audit(event_id, principal, action, path_hash, outcome)

Command and event envelopes

# Command envelope
operation: POST /v1/secrets/{path}/versions
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: secrets-management_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: secrets-management_123
event_type: secrets_management_committed
schema_version: 3
trace_id: tr_8ac...

The authoritative aggregate and idempotency record share a transactional or conditional-write boundary. Event streams, search indexes, caches, analytics, and dashboards are rebuildable projections.

↑ top

4. Why this architecture and what it costs

The boxes are not a shopping list. Each boundary exists because it owns a different invariant, scaling dimension, data lifecycle, security policy, or failure domain. The table makes those choices explicit and shows when the design should remain simpler.

Component	Owned responsibility	Why separate it	Rejected alternative	Cost and evolution trigger
Workload Identity	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Authenticate workload identity and evaluate path/action policy plus approval constraints.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Secrets API.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “KMS unavailable during secret read” to the external caller.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Workload Identity boundary before extracting or scaling it.
Secrets API	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Fetch ciphertext and ask HSM/KMS to unwrap only the data key.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Workload Identity from also owning the contract and failure behavior of Auth / Policy.	Put protocol, authentication, quotas, and business transitions in one process. Acceptable at low scale, but it couples external traffic risk to domain releases. Here that simpler choice would couple “Leaked lease remains cached” to Workload Identity.	Costs another hop, policy distribution, and a high-blast-radius fleet. Introduce it when many clients or services need one protocol and protection boundary. Measure the latency, saturation, lag, or mismatch at the Secrets API boundary before extracting or scaling it.
Auth / Policy	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Return a versioned secret lease or mint dynamic credentials with bounded TTL.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Secrets API from also owning the contract and failure behavior of Approval Workflow.	Embed policy in every caller. Rules then drift by deployment version and emergency changes require fleet-wide redeployment. Here that simpler choice would couple “Rotation breaks one dependent” to Secrets API.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Auth / Policy boundary before extracting or scaling it.
Approval Workflow	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Client agent caches only within lease and renews before safety margin.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Auth / Policy from also owning the contract and failure behavior of HSM / KMS.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Policy grants wildcard path” to Auth / Policy.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Approval Workflow boundary before extracting or scaling it.
HSM / KMS	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Rotation publishes new version, updates dependents, and revokes old leases with audit.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Approval Workflow from also owning the contract and failure behavior of Envelope Crypto.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Break-glass access bypasses normal approval” to Approval Workflow.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the HSM / KMS boundary before extracting or scaling it.
Envelope Crypto	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Authenticate workload identity and evaluate path/action policy plus approval constraints.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents HSM / KMS from also owning the contract and failure behavior of Encrypted Secret Store.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “KMS unavailable during secret read” to HSM / KMS.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Envelope Crypto boundary before extracting or scaling it.
Encrypted Secret Store	Owns authoritative versioned state or immutable evidence. In this design it participates in: Fetch ciphertext and ask HSM/KMS to unwrap only the data key.	Defines the source of truth and commit boundary. It prevents Envelope Crypto from also owning the contract and failure behavior of Version Metadata.	Let each service persist its own interpretation. That removes one clear commit boundary and makes retries, support, and reconciliation ambiguous. Here that simpler choice would couple “Leaked lease remains cached” to Envelope Crypto.	Costs schema migration, replication, backup, and shard ownership. Split storage only when access patterns, retention, or compliance boundaries materially differ. Measure the latency, saturation, lag, or mismatch at the Encrypted Secret Store boundary before extracting or scaling it.
Version Metadata	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Return a versioned secret lease or mint dynamic credentials with bounded TTL.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Encrypted Secret Store from also owning the contract and failure behavior of Lease Issuer.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Rotation breaks one dependent” to Encrypted Secret Store.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Version Metadata boundary before extracting or scaling it.
Lease Issuer	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Client agent caches only within lease and renews before safety margin.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Version Metadata from also owning the contract and failure behavior of Dynamic Credential Plugin.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Policy grants wildcard path” to Version Metadata.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Lease Issuer boundary before extracting or scaling it.
Dynamic Credential Plugin	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Rotation publishes new version, updates dependents, and revokes old leases with audit.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Lease Issuer from also owning the contract and failure behavior of Rotation Scheduler.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Break-glass access bypasses normal approval” to Lease Issuer.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Dynamic Credential Plugin boundary before extracting or scaling it.
Rotation Scheduler	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Authenticate workload identity and evaluate path/action policy plus approval constraints.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Dynamic Credential Plugin from also owning the contract and failure behavior of Revocation Stream.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “KMS unavailable during secret read” to Dynamic Credential Plugin.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Rotation Scheduler boundary before extracting or scaling it.
Revocation Stream	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Fetch ciphertext and ask HSM/KMS to unwrap only the data key.	Decouples bursts and failure from user acknowledgment. It prevents Rotation Scheduler from also owning the contract and failure behavior of Client Agent / Cache.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Leaked lease remains cached” to Rotation Scheduler.	Costs eventual consistency, schema evolution, replay tooling, idempotent consumers, and lag operations. Earn it when bursts or downstream failure must not block acknowledgment. Measure the latency, saturation, lag, or mismatch at the Revocation Stream boundary before extracting or scaling it.
Client Agent / Cache	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Return a versioned secret lease or mint dynamic credentials with bounded TTL.	Removes repeated work from the latency-critical path. It prevents Revocation Stream from also owning the contract and failure behavior of Audit Log.	Read authoritative storage for every request. Simpler initially, but latency and source-of-truth load scale together and a hot key becomes a database incident. Here that simpler choice would couple “Rotation breaks one dependent” to Revocation Stream.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (Millions secrets · 100K reads/s · HSM-backed root keys) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Client Agent / Cache boundary before extracting or scaling it.
Audit Log	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Client agent caches only within lease and renews before safety margin.	Proves correctness and supports controlled recovery. It prevents Client Agent / Cache from also owning the contract and failure behavior of Break-Glass Control.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Policy grants wildcard path” to Client Agent / Cache.	Costs eventual consistency, schema evolution, replay tooling, idempotent consumers, and lag operations. Earn it when bursts or downstream failure must not block acknowledgment. Measure the latency, saturation, lag, or mismatch at the Audit Log boundary before extracting or scaling it.
Break-Glass Control	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Rotation publishes new version, updates dependents, and revokes old leases with audit.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Audit Log from also owning the contract and failure behavior of operators and repair tooling.	Keep this behavior inside an adjacent service. Prefer that simpler option until independent scale, ownership, policy, or failure isolation is measurable. Here that simpler choice would couple “Break-glass access bypasses normal approval” to Audit Log.	Costs an interface, deployment, telemetry, ownership, and failure mode. Keep it in-process first; extract when independent scale or reliability pays for the boundary. Measure the latency, saturation, lag, or mismatch at the Break-Glass Control boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Use envelope encryption so the database never holds plaintext or root keys; HSM calls protect key operations rather than every application byte.

A single deployable with one transactional store is the right starting point for a small secrets management workload. The proposed boundary becomes justified when use envelope encryption so the database never holds plaintext or root keys; hsm calls protect key operations rather than every application byte. Separating it contains the named failure “KMS unavailable during secret read” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Short-lived dynamic credentials reduce blast radius but introduce lease renewal and dependency availability costs.

The design coordinates only where each accepted secrets management operation has one stable result and retries do not repeat its business effect. Everywhere else it accepts bounded staleness or asynchronous propagation. The alternative—strong coordination for every projection—simplifies the mental model but increases tail latency and turns partitions into broad write unavailability.

Decision 03

Operational cost and evolution

Authorization is identity/resource/action-specific and every read is audited; caching plaintext broadly defeats the service boundary.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Return a versioned secret lease or mint dynamic credentials with bounded TTL.” and extract another service only when measured throughput, blast radius, data lifecycle, or team ownership changes independently.

Why this design: the synchronous path contains only the decisions required to protect each accepted secrets management operation has one stable result and retries do not repeat its business effect. Durable projections, notifications, indexing, analytics, and repair are isolated so their load or failure cannot corrupt the acknowledged operation.

Correctness boundary

Use envelope encryption so the database never holds plaintext or root keys; HSM calls protect key operations rather than every application byte.

Scale boundary

Short-lived dynamic credentials reduce blast radius but introduce lease renewal and dependency availability costs.

Operational boundary

Authorization is identity/resource/action-specific and every read is audited; caching plaintext broadly defeats the service boundary.

↑ top

5. Critical flow and algorithm

Authenticate workload identity and evaluate path/action policy plus approval constraints.

Fetch ciphertext and ask HSM/KMS to unwrap only the data key.

Return a versioned secret lease or mint dynamic credentials with bounded TTL.

Client agent caches only within lease and renews before safety margin.

Rotation publishes new version, updates dependents, and revokes old leases with audit.

Retry-safe mutation skeleton

identity = authenticate(request)
prior = idempotency.get(identity, request.key)
if prior: return prior.result

current = authoritative_store.read(partition_key)
next_state, events = domain.transition(current, request)
authoritative_store.compare_and_set(current.version, next_state)
idempotency.commit(request.key, next_state.result)
outbox.append(events)
return next_state.result

Exactly-once business effect: the network can duplicate requests. Stable identities, conditional state transitions, immutable event IDs, and reconciliation make the user-visible effect happen once.

↑ top

6. Caching, hot keys, and backpressure

Cache policy

Cache immutable or slowly changing reads with versioned keys, TTL jitter, short negative entries, and request coalescing.

Hot partition or tenant

Measure per-key and per-tenant skew. Replicate safe reads, isolate the offender, salt only when ordering permits, and enforce budgets.

Backpressure

Bound queues, buffers, connection pools, and concurrency. Shed low-priority work and scale from lag/oldest-item age rather than CPU alone.

Stampede and retry storms

Use single-flight/coalescing, jittered expiry, bounded retries, exponential backoff, and circuit breakers so one dependency failure does not multiply traffic.

Cache outage rule: never allow every miss to fall through simultaneously. Use local protection, request coalescing, circuit breakers, bounded pools, and controlled fallback.

↑ top

7. Partitioning and consistency

partition = stable_hash(primary_or_aggregate_key) → shard owner
write = conditional_update(expected_version)
read_projection = versioned and rebuildable

Strong boundary

each accepted secrets management operation has one stable result and retries do not repeat its business effect. Use a uniqueness constraint, serializable/conditional write, expected version, or fencing token at this boundary.

Eventual boundary

Analytics, notifications, search, read models, and operational dashboards may lag when freshness is observable and replay exists.

Operation	Consistency	Reason
Authoritative mutation	Conditional / serializable	Protect uniqueness, ownership, money, inventory, order, or legal state transition.
Read after own write	Primary, session token, or cache warm	The caller must observe the result it just committed.
Ordinary read projection	Bounded stale where safe	Replicas and caches improve latency when version/freshness is visible.
Analytics and operations	Eventual + replay	Freshness is less important than isolating the online path.

Route every mutation for one aggregate to one owner/partition epoch.
Use cursors rather than offsets for changing ordered collections.
Retain tombstones until replication and repair windows close.
Move shard ownership through versioned maps and dual-read verification.

↑ top

8. Failure drills and recovery

Failure	Runtime behavior	Recovery proof
KMS unavailable during secret read	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Leaked lease remains cached	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Rotation breaks one dependent	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Policy grants wildcard path	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Break-glass access bypasses normal approval	Stop unsafe progress, serve only explicitly safe degraded behavior, and recover from durable state.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.

Recovery checkpoint

# Recovery checkpoint
scenario: KMS unavailable during secret read
operation_id: op_01J...
aggregate_key: secrets-management_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Break-Glass Control

Every dependency needs a timeout, retry budget, circuit policy, and explicit fail-open, fail-closed, stale, or read-only mode. “Highly available” is incomplete until degraded behavior is named.

↑ top

9. Security, regions, and observability

Security and abuse

Authenticate at ingress and authorize the exact resource/tenant on every operation.
Encrypt transport and durable data; tokenize or isolate sensitive material.
Rate-limit by tenant, actor, resource, and network; detect enumeration and anomalous fan-out.
Redact secrets and personal data before logs or analytical storage.
Audit privileged mutations, support access, policy changes, and forced recovery.

Signals that prove correctness

User-visible success rate and p50/p95/p99 latency.
Timeouts, circuit state, saturation, and replica health by dependency.
Queue lag, oldest-item age, retries, and dead-letter growth.
Conflict, duplicate-suppression, and reconciliation mismatch counts.
Business drill: KMS unavailable during secret read — alert before the user-visible invariant is violated.

Regional evolution

Begin multi-zone with tested failover, backups, and reconciliation.
Add global stateless ingress and regional read projections.
Keep one write owner per aggregate or partition where coordination is expensive.
Move ownership with epochs/fencing; avoid silent simultaneous writers.
Use expand/contract schemas, canaries, shadow reads, and reversible cutovers.

↑ top

10. Interviewer follow-ups

How does envelope encryption work?

Use envelope encryption so the database never holds plaintext or root keys; HSM calls protect key operations rather than every application byte. Walk the concrete lifecycle step “Authenticate workload identity and evaluate path/action policy plus approval constraints.” Then test it against “KMS unavailable during secret read”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are secrets rotated?

Short-lived dynamic credentials reduce blast radius but introduce lease renewal and dependency availability costs. Walk the concrete lifecycle step “Fetch ciphertext and ask HSM/KMS to unwrap only the data key.” Then test it against “Leaked lease remains cached”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

What if KMS is unavailable?

Authorization is identity/resource/action-specific and every read is audited; caching plaintext broadly defeats the service boundary. Walk the concrete lifecycle step “Return a versioned secret lease or mint dynamic credentials with bounded TTL.” Then test it against “Rotation breaks one dependent”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do dynamic database credentials work?

Use envelope encryption so the database never holds plaintext or root keys; HSM calls protect key operations rather than every application byte. Walk the concrete lifecycle step “Client agent caches only within lease and renews before safety margin.” Then test it against “Policy grants wildcard path”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do applications bootstrap identity?

Short-lived dynamic credentials reduce blast radius but introduce lease renewal and dependency availability costs. Walk the concrete lifecycle step “Rotation publishes new version, updates dependents, and revokes old leases with audit.” Then test it against “Break-glass access bypasses normal approval”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

Answer pattern: state the invariant, choose a mechanism, walk one failure, name the recovery signal, and explain the cost of the alternative.

↑ top

11. How to present this in 45 minutes

Time	What to cover
0–5 min	Clarify scope, workload, SLOs, and the correctness invariant.
5–10 min	Estimate throughput/storage and define concrete APIs and records.
10–20 min	Draw online, durable-write, and asynchronous/repair paths.
20–30 min	Explain the exact lifecycle, partition key, and concurrency control.
30–38 min	Discuss caching, hot keys, replication, and backpressure.
38–43 min	Run failure drills, security, observability, and regional evolution.
43–45 min	Summarize the five most important choices and trade-offs.

Final decision summary

Invariant	each accepted secrets management operation has one stable result and retries do not repeat its business effect.
Partitioning	Stable aggregate/key hash with versioned ownership and measured hot-key isolation.
Commit	Expected version or uniqueness at the authoritative state boundary.
Async work	Durable outbox/log, idempotent consumers, checkpoints, and replay.
Recovery	Same operation identity, durable truth lookup, and business reconciliation.

Closing summary

CLOSING “The design protects each accepted secrets management operation has one stable result and retries do not repeat its business effect. The online path stays bounded, authoritative state is versioned and partitioned, and asynchronous work is replayable and observable.”

↑ top

Workload Identity

Secrets API

Auth / Policy

Approval Workflow

Complete System Design Tutorial

1. Requirements and scope

Functional requirements

Non-functional requirements

2. Capacity assumptions

3. API and data model

Authoritative records and indexes

Command and event envelopes

4. Why this architecture and what it costs

Chosen boundary vs simpler monolith

Correctness vs availability

Operational cost and evolution

5. Critical flow and algorithm

Retry-safe mutation skeleton

6. Caching, hot keys, and backpressure

7. Partitioning and consistency

Strong boundary

Eventual boundary

8. Failure drills and recovery

Recovery checkpoint

9. Security, regions, and observability

Security and abuse

Signals that prove correctness

Regional evolution

10. Interviewer follow-ups

11. How to present this in 45 minutes

Final decision summary

Closing summary