Logging Framework — 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 logging framework 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 logs/s in-process · bounded memory · pluggable sinks.
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: every object transition is legal and atomic at one aggregate boundary.

↑ top

2. Capacity assumptions

State assumptions before choosing technology. The supplied workload is Millions logs/s in-process · bounded memory · pluggable sinks. 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 logs/s in-process · bounded memory · pluggable sinks	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

logger.info(template, fields)structured event

with_context(trace_id, tenant)scoped metadata

set_level(logger, version)dynamic config

flush(deadline)bounded shutdown

Authoritative records and indexes

LogEvent(timestamp, level, logger, template, fields, context, throwable)

Appender(name, filter, formatter, failure_policy)

RotationPolicy(max_bytes, interval, retained_files)

BufferState(capacity, used, dropped_by_level)

Command and event envelopes

# Command envelope
operation: logger.info(template, fields)
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: logging-framework_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: logging-framework_123
event_type: logging_framework_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
Application Call Site	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Check level before allocating expensive event fields.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Logger Facade.	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 “Network appender recursively logs failure” 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 Application Call Site boundary before extracting or scaling it.
Logger Facade	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Build immutable structured event and redact known sensitive keys at the boundary.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Application Call Site from also owning the contract and failure behavior of Level Filter.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Buffer fills during outage” to Application Call Site.	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 Logger Facade boundary before extracting or scaling it.
Level Filter	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Offer to bounded ring buffer under explicit block/drop/sample policy by severity.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Logger Facade from also owning the contract and failure behavior of Context / MDC.	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 “Mutable context leaks across requests” to Logger Facade.	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 Level Filter boundary before extracting or scaling it.
Context / MDC	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Dispatcher batches, formats once per appender needs, and isolates sink failures.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Level Filter from also owning the contract and failure behavior of Event Builder.	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 loses file handle” to Level Filter.	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 Context / MDC boundary before extracting or scaling it.
Event Builder	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Rotation atomically renames/reopens; shutdown flushes until deadline then records drops.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Context / MDC from also owning the contract and failure behavior of Redaction Filter.	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 “Shutdown deadlocks application thread” to Context / MDC.	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 Event Builder boundary before extracting or scaling it.
Redaction Filter	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Check level before allocating expensive event fields.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Event Builder from also owning the contract and failure behavior of Bounded Ring Buffer.	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 “Network appender recursively logs failure” to Event Builder.	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 Redaction Filter boundary before extracting or scaling it.
Bounded Ring Buffer	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Build immutable structured event and redact known sensitive keys at the boundary.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Redaction Filter from also owning the contract and failure behavior of Async Dispatcher.	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 “Buffer fills during outage” to Redaction Filter.	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 Bounded Ring Buffer boundary before extracting or scaling it.
Async Dispatcher	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Offer to bounded ring buffer under explicit block/drop/sample policy by severity.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Bounded Ring Buffer from also owning the contract and failure behavior of Formatter.	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 “Mutable context leaks across requests” to Bounded Ring Buffer.	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 Async Dispatcher boundary before extracting or scaling it.
Formatter	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Dispatcher batches, formats once per appender needs, and isolates sink failures.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Async Dispatcher from also owning the contract and failure behavior of Console Appender.	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 loses file handle” to Async Dispatcher.	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 Formatter boundary before extracting or scaling it.
Console Appender	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Rotation atomically renames/reopens; shutdown flushes until deadline then records drops.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Formatter from also owning the contract and failure behavior of File Appender.	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 “Shutdown deadlocks application thread” to Formatter.	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 Console Appender boundary before extracting or scaling it.
File Appender	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Check level before allocating expensive event fields.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Console Appender from also owning the contract and failure behavior of Network Appender.	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 “Network appender recursively logs failure” to Console Appender.	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 File Appender boundary before extracting or scaling it.
Network Appender	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Build immutable structured event and redact known sensitive keys at the boundary.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents File Appender from also owning the contract and failure behavior of Rotation Manager.	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 “Buffer fills during outage” to File Appender.	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 Network Appender boundary before extracting or scaling it.
Rotation Manager	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Offer to bounded ring buffer under explicit block/drop/sample policy by severity.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Network Appender from also owning the contract and failure behavior of Failure Fallback.	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 “Mutable context leaks across requests” to Network Appender.	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 Manager boundary before extracting or scaling it.
Failure Fallback	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Dispatcher batches, formats once per appender needs, and isolates sink failures.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Rotation Manager from also owning the contract and failure behavior of Metrics.	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 loses file handle” to Rotation Manager.	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 Failure Fallback boundary before extracting or scaling it.
Metrics	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Rotation atomically renames/reopens; shutdown flushes until deadline then records drops.	Proves correctness and supports controlled recovery. It prevents Failure Fallback 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 “Shutdown deadlocks application thread” to Failure Fallback.	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 Metrics boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Separate log event creation, filtering, formatting, and transport through small interfaces.

A single deployable with one transactional store is the right starting point for a small logging framework workload. The proposed boundary becomes justified when separate log event creation, filtering, formatting, and transport through small interfaces. Separating it contains the named failure “Network appender recursively logs failure” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Use a bounded async queue with an explicit drop/block policy; logging must not deadlock the application.

The design coordinates only where every object transition is legal and atomic at one aggregate boundary. 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

Carry structured context and trace IDs without mutable global maps; sanitize sensitive fields.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Offer to bounded ring buffer under explicit block/drop/sample policy by severity.” 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 every object transition is legal and atomic at one aggregate boundary. Durable projections, notifications, indexing, analytics, and repair are isolated so their load or failure cannot corrupt the acknowledged operation.

Correctness boundary

Separate log event creation, filtering, formatting, and transport through small interfaces.

Scale boundary

Use a bounded async queue with an explicit drop/block policy; logging must not deadlock the application.

Operational boundary

Carry structured context and trace IDs without mutable global maps; sanitize sensitive fields.

↑ top

5. Critical flow and algorithm

Check level before allocating expensive event fields.

Build immutable structured event and redact known sensitive keys at the boundary.

Offer to bounded ring buffer under explicit block/drop/sample policy by severity.

Dispatcher batches, formats once per appender needs, and isolates sink failures.

Rotation atomically renames/reopens; shutdown flushes until deadline then records drops.

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

Add a cache only for a measured repeated read; do not place mutable correctness state in a disposable layer.

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

every object transition is legal and atomic at one aggregate boundary. 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
Network appender recursively logs failure	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.
Buffer fills during outage	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.
Mutable context leaks across requests	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 loses file handle	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.
Shutdown deadlocks application thread	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: Network appender recursively logs failure
operation_id: op_01J...
aggregate_key: logging-framework_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Metrics

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: Network appender recursively logs failure — 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 do async appenders handle shutdown?

Separate log event creation, filtering, formatting, and transport through small interfaces. Walk the concrete lifecycle step “Check level before allocating expensive event fields.” Then test it against “Network appender recursively logs failure”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How does log rotation work?

Use a bounded async queue with an explicit drop/block policy; logging must not deadlock the application. Walk the concrete lifecycle step “Build immutable structured event and redact known sensitive keys at the boundary.” Then test it against “Buffer fills during outage”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are dynamic levels updated?

Carry structured context and trace IDs without mutable global maps; sanitize sensitive fields. Walk the concrete lifecycle step “Offer to bounded ring buffer under explicit block/drop/sample policy by severity.” Then test it against “Mutable context leaks across requests”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you avoid recursive logging failures?

Separate log event creation, filtering, formatting, and transport through small interfaces. Walk the concrete lifecycle step “Dispatcher batches, formats once per appender needs, and isolates sink failures.” Then test it against “Rotation loses file handle”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you test backpressure?

Use a bounded async queue with an explicit drop/block policy; logging must not deadlock the application. Walk the concrete lifecycle step “Rotation atomically renames/reopens; shutdown flushes until deadline then records drops.” Then test it against “Shutdown deadlocks application thread”. 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	every object transition is legal and atomic at one aggregate boundary.
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 every object transition is legal and atomic at one aggregate boundary. The online path stays bounded, authoritative state is versioned and partitioned, and asynchronous work is replayable and observable.”

↑ top

Application Call Site

Logger Facade

Level Filter

Context / MDC

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