CI/CD Platform — 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 ci/cd platform 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 1M builds/day · untrusted code · petabytes of artifacts.
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 ci/cd platform 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 1M builds/day · untrusted code · petabytes of artifacts. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	1M builds/day · untrusted code · petabytes of artifacts	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/hooks/gitsigned event

POST /v1/pipelines/{id}/runs

POST /v1/runs/{id}/cancel

POST /v1/deployments/{id}/promote

Authoritative records and indexes

pipeline(id, repo, definition_version, policy)

run(id PK, commit, state, dag_version)

task(run_id, node_id, attempt, lease, state)

artifact(digest PK, uri, provenance, retention)

Command and event envelopes

# Command envelope
operation: POST /v1/hooks/git
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: ci-cd-platform_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: ci-cd-platform_123
event_type: ci_cd_platform_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
Git Provider	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Verify webhook and compile pinned pipeline definition into immutable DAG.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Webhook Ingest.	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 “Malicious build escapes sandbox” 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 Git Provider boundary before extracting or scaling it.
Webhook Ingest	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Scheduler releases ready tasks under tenant quota/priority and leases ephemeral runners.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Git Provider from also owning the contract and failure behavior of Pipeline 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 “Runner dies after artifact upload” to Git Provider.	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 Webhook Ingest boundary before extracting or scaling it.
Pipeline API	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Runner fetches source/cache, executes in sandbox, and streams bounded logs.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Webhook Ingest from also owning the contract and failure behavior of Definition Store.	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 “Cache poisoning crosses repositories” to Webhook Ingest.	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 Pipeline API boundary before extracting or scaling it.
Definition Store	Owns authoritative versioned state or immutable evidence. In this design it participates in: Publish content-addressed artifacts plus provenance before marking task success.	Defines the source of truth and commit boundary. It prevents Pipeline API from also owning the contract and failure behavior of Workflow Compiler.	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 “Webhook replay duplicates pipeline” to Pipeline API.	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 Definition Store boundary before extracting or scaling it.
Workflow Compiler	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Deployment controller promotes exact artifacts through approvals, canaries, and rollback.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Definition Store from also owning the contract and failure behavior of DAG 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 “Deployment health gate reports late” to Definition 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 Workflow Compiler boundary before extracting or scaling it.
DAG Scheduler	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Verify webhook and compile pinned pipeline definition into immutable DAG.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Workflow Compiler from also owning the contract and failure behavior of Quota / Priority.	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 “Malicious build escapes sandbox” to Workflow Compiler.	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 DAG Scheduler boundary before extracting or scaling it.
Quota / Priority	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Scheduler releases ready tasks under tenant quota/priority and leases ephemeral runners.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents DAG Scheduler from also owning the contract and failure behavior of Ephemeral Runner.	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 “Runner dies after artifact upload” to DAG Scheduler.	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 Quota / Priority boundary before extracting or scaling it.
Ephemeral Runner	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Runner fetches source/cache, executes in sandbox, and streams bounded logs.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Quota / Priority from also owning the contract and failure behavior of Sandbox / Network Policy.	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 “Cache poisoning crosses repositories” to Quota / Priority.	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 Ephemeral Runner boundary before extracting or scaling it.
Sandbox / Network Policy	Owns authoritative versioned state or immutable evidence. In this design it participates in: Publish content-addressed artifacts plus provenance before marking task success.	Defines the source of truth and commit boundary. It prevents Ephemeral Runner from also owning the contract and failure behavior of Source Cache.	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 “Webhook replay duplicates pipeline” to Ephemeral Runner.	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 Sandbox / Network Policy boundary before extracting or scaling it.
Source Cache	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Deployment controller promotes exact artifacts through approvals, canaries, and rollback.	Removes repeated work from the latency-critical path. It prevents Sandbox / Network Policy from also owning the contract and failure behavior of Build Cache.	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 “Deployment health gate reports late” to Sandbox / Network Policy.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1M builds/day · untrusted code · petabytes of artifacts) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Source Cache boundary before extracting or scaling it.
Build Cache	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Verify webhook and compile pinned pipeline definition into immutable DAG.	Removes repeated work from the latency-critical path. It prevents Source Cache from also owning the contract and failure behavior of Artifact Store.	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 “Malicious build escapes sandbox” to Source Cache.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (1M builds/day · untrusted code · petabytes of artifacts) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Build Cache boundary before extracting or scaling it.
Artifact Store	Owns authoritative versioned state or immutable evidence. In this design it participates in: Scheduler releases ready tasks under tenant quota/priority and leases ephemeral runners.	Defines the source of truth and commit boundary. It prevents Build Cache from also owning the contract and failure behavior of Log Stream.	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 “Runner dies after artifact upload” to Build Cache.	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 Artifact Store boundary before extracting or scaling it.
Log Stream	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Runner fetches source/cache, executes in sandbox, and streams bounded logs.	Decouples bursts and failure from user acknowledgment. It prevents Artifact Store from also owning the contract and failure behavior of Deployment Controller.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Cache poisoning crosses repositories” to Artifact Store.	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 Log Stream boundary before extracting or scaling it.
Deployment Controller	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Publish content-addressed artifacts plus provenance before marking task success.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Log Stream from also owning the contract and failure behavior of Audit / Provenance.	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 “Webhook replay duplicates pipeline” to Log Stream.	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 Deployment Controller boundary before extracting or scaling it.
Audit / Provenance	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Deployment controller promotes exact artifacts through approvals, canaries, and rollback.	Proves correctness and supports controlled recovery. It prevents Deployment Controller from also owning the contract and failure behavior of operators and repair tooling.	Infer correctness from request logs. Logs do not prove durable business state, detect silent drift, or safely repair ambiguous outcomes. Here that simpler choice would couple “Deployment health gate reports late” to Deployment Controller.	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 Audit / Provenance boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Compile declarative pipelines into immutable DAG executions; dynamic ad-hoc orchestration makes retries and audit impossible.

A single deployable with one transactional store is the right starting point for a small ci/cd platform workload. The proposed boundary becomes justified when compile declarative pipelines into immutable dag executions; dynamic ad-hoc orchestration makes retries and audit impossible. Separating it contains the named failure “Malicious build escapes sandbox” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Run untrusted builds in ephemeral sandboxes with scoped credentials and network policy; sharing warm workers trades speed for isolation risk.

The design coordinates only where each accepted ci/cd platform 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

Artifacts and logs are immutable outputs, while deployment promotion is a separate, policy-gated state transition.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Runner fetches source/cache, executes in sandbox, and streams bounded logs.” 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 ci/cd platform 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

Compile declarative pipelines into immutable DAG executions; dynamic ad-hoc orchestration makes retries and audit impossible.

Scale boundary

Run untrusted builds in ephemeral sandboxes with scoped credentials and network policy; sharing warm workers trades speed for isolation risk.

Operational boundary

Artifacts and logs are immutable outputs, while deployment promotion is a separate, policy-gated state transition.

↑ top

5. Critical flow and algorithm

Verify webhook and compile pinned pipeline definition into immutable DAG.

Scheduler releases ready tasks under tenant quota/priority and leases ephemeral runners.

Runner fetches source/cache, executes in sandbox, and streams bounded logs.

Publish content-addressed artifacts plus provenance before marking task success.

Deployment controller promotes exact artifacts through approvals, canaries, and rollback.

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 ci/cd platform 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
Malicious build escapes sandbox	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.
Runner dies after artifact upload	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.
Cache poisoning crosses repositories	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.
Webhook replay duplicates pipeline	Resolve at source of truth with uniqueness, expected version, lease generation, or monotonic cursor.	Reconciliation returns zero invariant gaps; lag/replica health and user-visible SLO return to baseline.
Deployment health gate reports late	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: Malicious build escapes sandbox
operation_id: op_01J...
aggregate_key: ci-cd-platform_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Audit / Provenance

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: Malicious build escapes sandbox — 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 are runners isolated?

Compile declarative pipelines into immutable DAG executions; dynamic ad-hoc orchestration makes retries and audit impossible. Walk the concrete lifecycle step “Verify webhook and compile pinned pipeline definition into immutable DAG.” Then test it against “Malicious build escapes sandbox”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does build caching work?

Run untrusted builds in ephemeral sandboxes with scoped credentials and network policy; sharing warm workers trades speed for isolation risk. Walk the concrete lifecycle step “Scheduler releases ready tasks under tenant quota/priority and leases ephemeral runners.” Then test it against “Runner dies after artifact upload”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do retries preserve semantics?

Artifacts and logs are immutable outputs, while deployment promotion is a separate, policy-gated state transition. Walk the concrete lifecycle step “Runner fetches source/cache, executes in sandbox, and streams bounded logs.” Then test it against “Cache poisoning crosses repositories”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are secrets injected?

Compile declarative pipelines into immutable DAG executions; dynamic ad-hoc orchestration makes retries and audit impossible. Walk the concrete lifecycle step “Publish content-addressed artifacts plus provenance before marking task success.” Then test it against “Webhook replay duplicates pipeline”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do progressive deployments work?

Run untrusted builds in ephemeral sandboxes with scoped credentials and network policy; sharing warm workers trades speed for isolation risk. Walk the concrete lifecycle step “Deployment controller promotes exact artifacts through approvals, canaries, and rollback.” Then test it against “Deployment health gate reports late”. 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 ci/cd platform 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 ci/cd platform 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

Git Provider

Webhook Ingest

Pipeline API

Definition Store

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