Logging & Monitoring — 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 & monitoring 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 20 TB/day · million events/s · seconds-to-search.
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 logging & monitoring 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 20 TB/day · million events/s · seconds-to-search. Convert it into peak throughput, retained records, bandwidth, and failure budget; then connect each number to a component.

Dimension	Working estimate	Design implication
Traffic	20 TB/day · million events/s · seconds-to-search	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/logscompressed batch ingest

GET /v1/search?query=&cursor=tenant query

POST /v1/rulesversioned alert rule

GET /v1/export/{job}archived result

Authoritative records and indexes

log_event(tenant, event_id, timestamp, service, severity, fields)

segment(tenant, time_range, index_uri, data_uri, checksum)

alert_rule(rule_id, query, window, threshold, version)

tenant_budget(tenant, bytes_per_sec, retention_class)

Command and event envelopes

# Command envelope
operation: POST /v1/logs
request_id: req_01J...
idempotency_key: idem_7f2...
aggregate_key: logging-monitoring_123
expected_version: 17
payload: {...}

# Durable event
event_id: evt_01J...
aggregate_key: logging-monitoring_123
event_type: logging_monitoring_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
App SDK / Agent	Owns client protocol, request identity, retry, and resume behavior. In this design it participates in: Batch/compress at the agent and spill to bounded disk when collectors fail.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents the external caller from also owning the contract and failure behavior of Local Spool.	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 “Logging storm consumes service disk” 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 App SDK / Agent boundary before extracting or scaling it.
Local Spool	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Collectors authenticate tenant, enforce byte/cardinality quotas, and append durably.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents App SDK / Agent from also owning the contract and failure behavior of Regional Collector.	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 “Parser poison event blocks partition” to App SDK / Agent.	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 Local Spool boundary before extracting or scaling it.
Regional Collector	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Parse against a schema, redact secrets, and route time-partitioned segments to indexers.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Local Spool from also owning the contract and failure behavior of Tenant Quota.	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 “Tenant query fans across every shard” to Local Spool.	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 Regional Collector boundary before extracting or scaling it.
Tenant Quota	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Serve recent indexed search; federate archived scans asynchronously for old windows.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Regional Collector from also owning the contract and failure behavior of Durable Ingest Log.	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 “Secret reaches archive before redaction” to Regional Collector.	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 Tenant Quota boundary before extracting or scaling it.
Durable Ingest Log	Preserves durable ordering, replay, and consumer checkpoints. In this design it participates in: Evaluate alert windows from durable offsets and dedupe notifications by incident key.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Tenant Quota from also owning the contract and failure behavior of Parser / Redactor.	Call every downstream consumer synchronously. That couples user latency and availability to analytics, notification, indexing, and audit dependencies. Here that simpler choice would couple “Alert evaluator replays and duplicates pages” to Tenant Quota.	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 Durable Ingest Log boundary before extracting or scaling it.
Parser / Redactor	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Batch/compress at the agent and spill to bounded disk when collectors fail.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Durable Ingest Log from also owning the contract and failure behavior of Schema Registry.	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 “Logging storm consumes service disk” to Durable Ingest 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 Parser / Redactor boundary before extracting or scaling it.
Schema Registry	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Collectors authenticate tenant, enforce byte/cardinality quotas, and append durably.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Parser / Redactor from also owning the contract and failure behavior of Hot Indexer.	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 “Parser poison event blocks partition” to Parser / Redactor.	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 Schema Registry boundary before extracting or scaling it.
Hot Indexer	Serves a low-latency, disposable projection with authoritative fallback. In this design it participates in: Parse against a schema, redact secrets, and route time-partitioned segments to indexers.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Schema Registry from also owning the contract and failure behavior of Search Shards.	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 “Tenant query fans across every shard” to Schema Registry.	Costs invalidation, staleness, stampede protection, and memory. Add or split it only after the workload (20 TB/day · million events/s · seconds-to-search) proves repeated reads dominate. Measure the latency, saturation, lag, or mismatch at the Hot Indexer boundary before extracting or scaling it.
Search Shards	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Serve recent indexed search; federate archived scans asynchronously for old windows.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Hot Indexer from also owning the contract and failure behavior of Trace Correlator.	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 “Secret reaches archive before redaction” to Hot Indexer.	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 Search Shards boundary before extracting or scaling it.
Trace Correlator	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Evaluate alert windows from durable offsets and dedupe notifications by incident key.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Search Shards from also owning the contract and failure behavior of Alert Evaluator.	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 “Alert evaluator replays and duplicates pages” to Search Shards.	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 Trace Correlator boundary before extracting or scaling it.
Alert Evaluator	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Batch/compress at the agent and spill to bounded disk when collectors fail.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Trace Correlator from also owning the contract and failure behavior of Object Archive.	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 “Logging storm consumes service disk” to Trace Correlator.	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 Alert Evaluator boundary before extracting or scaling it.
Object Archive	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Collectors authenticate tenant, enforce byte/cardinality quotas, and append durably.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Alert Evaluator from also owning the contract and failure behavior of Query Gateway.	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 “Parser poison event blocks partition” to Alert Evaluator.	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 Object Archive boundary before extracting or scaling it.
Query Gateway	Owns authentication, validation, quotas, deadlines, and routing. In this design it participates in: Parse against a schema, redact secrets, and route time-partitioned segments to indexers.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Object Archive from also owning the contract and failure behavior of Retention Compactor.	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 “Tenant query fans across every shard” to Object Archive.	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 Query Gateway boundary before extracting or scaling it.
Retention Compactor	Owns one bounded domain decision, transition, or operational capability. In this design it participates in: Serve recent indexed search; federate archived scans asynchronously for old windows.	Separates a distinct scale, ownership, policy, or failure boundary. It prevents Query Gateway 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 “Secret reaches archive before redaction” to Query Gateway.	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 Retention Compactor boundary before extracting or scaling it.

Decision 01

Chosen boundary vs simpler monolith

Buffer before parsing; collectors must apply quotas and backpressure instead of exhausting memory.

A single deployable with one transactional store is the right starting point for a small logging & monitoring workload. The proposed boundary becomes justified when buffer before parsing; collectors must apply quotas and backpressure instead of exhausting memory. Separating it contains the named failure “Logging storm consumes service disk” without placing unrelated work on the critical path.

Decision 02

Correctness vs availability

Route recent searchable data to hot storage and compact older data into object storage.

The design coordinates only where each accepted logging & monitoring 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

Control cardinality, redact secrets, and preserve tenant boundaries from ingest through query.

This choice adds dashboards, ownership, schema compatibility, capacity planning, and recovery runbooks. Start with the minimum topology that supports “Parse against a schema, redact secrets, and route time-partitioned segments to indexers.” 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 logging & monitoring 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

Buffer before parsing; collectors must apply quotas and backpressure instead of exhausting memory.

Scale boundary

Route recent searchable data to hot storage and compact older data into object storage.

Operational boundary

Control cardinality, redact secrets, and preserve tenant boundaries from ingest through query.

↑ top

5. Critical flow and algorithm

Batch/compress at the agent and spill to bounded disk when collectors fail.

Collectors authenticate tenant, enforce byte/cardinality quotas, and append durably.

Parse against a schema, redact secrets, and route time-partitioned segments to indexers.

Serve recent indexed search; federate archived scans asynchronously for old windows.

Evaluate alert windows from durable offsets and dedupe notifications by incident key.

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 logging & monitoring 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
Logging storm consumes service disk	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.
Parser poison event blocks partition	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.
Tenant query fans across every shard	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.
Secret reaches archive before redaction	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.
Alert evaluator replays and duplicates pages	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.

Recovery checkpoint

# Recovery checkpoint
scenario: Logging storm consumes service disk
operation_id: op_01J...
aggregate_key: logging-monitoring_123
last_durable_version: 18
projection_checkpoint: partition-07@offset-482901
retry_policy: same_identity / bounded_backoff
reconciliation_status: pending
owner: Retention Compactor

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: Logging storm consumes service disk — 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 you survive a logging storm?

Buffer before parsing; collectors must apply quotas and backpressure instead of exhausting memory. Walk the concrete lifecycle step “Batch/compress at the agent and spill to bounded disk when collectors fail.” Then test it against “Logging storm consumes service disk”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How are retention tiers implemented?

Route recent searchable data to hot storage and compact older data into object storage. Walk the concrete lifecycle step “Collectors authenticate tenant, enforce byte/cardinality quotas, and append durably.” Then test it against “Parser poison event blocks partition”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do alerts avoid duplicates?

Control cardinality, redact secrets, and preserve tenant boundaries from ingest through query. Walk the concrete lifecycle step “Parse against a schema, redact secrets, and route time-partitioned segments to indexers.” Then test it against “Tenant query fans across every shard”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you search archived logs?

Buffer before parsing; collectors must apply quotas and backpressure instead of exhausting memory. Walk the concrete lifecycle step “Serve recent indexed search; federate archived scans asynchronously for old windows.” Then test it against “Secret reaches archive before redaction”. State the retry identity, durable checkpoint, recovery signal, and cost of the alternative.

How do you prevent secret leakage?

Route recent searchable data to hot storage and compact older data into object storage. Walk the concrete lifecycle step “Evaluate alert windows from durable offsets and dedupe notifications by incident key.” Then test it against “Alert evaluator replays and duplicates pages”. 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 logging & monitoring 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 logging & monitoring 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

App SDK / Agent

Local Spool

Regional Collector

Tenant Quota

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