System Design: Designing a Distributed Logging System (TB/Day Scale)

Case Study: Design a Logging & Monitoring System

Mental Model

Connecting isolated components into a resilient, scalable, and observable distributed web.

Building a logging system for a single app is easy. Building one for 1,000 microservices generating terabytes of logs per hour is a High-Throughput Write problem.

In a microservices architecture with thousands of containers, logs are scattered everywhere. You need a centralized system that can ingest terabytes of log data every day, store it cost-effectively, and allow engineers to search it in near real-time.

1. Core Requirements

graph LR
    Producer[Producer Service] -->|Publish Event| Kafka[Kafka / Event Bus]
    Kafka -->|Consume| Consumer1[Consumer Group A]
    Kafka -->|Consume| Consumer2[Consumer Group B]
    Consumer1 --> DB1[(Primary DB)]
    Consumer2 --> Cache[(Redis)]

• High Throughput: Ingesting millions of log lines per second.
• Searchability: Full-text search across logs (errors, request IDs).
• Retention: Keeping "hot" logs for 7 days and "cold" logs for 1 year.
• Resilience: If the logging system is slow, it should not crash the main application.

2. The Ingestion Pipeline (The ELK Model)

The industry standard for logging is the ELK Stack (Elasticsearch, Logstash, Kibana).

Phase 1: The Collector (Filebeat/Fluentd)

A lightweight agent (daemonset) runs on every server/container. It monitors log files and pushes them to the next stage.

• Why? It ensures that if the network is down, logs are buffered locally on disk.

Phase 2: The Buffer (Apache Kafka)

You should never push logs directly to your database.

• The Problem: A spike in application traffic will create a spike in logs, which could overwhelm your search engine.
• The Solution: Use Kafka as a buffer. The collectors push to Kafka, and the indexing workers consume at a steady, sustainable rate.

Phase 3: The Transformer (Logstash)

Logstash (or a custom Flink job) pulls logs from Kafka, parses them (JSON, Grok), and enriches them (adding region_id or user_metadata).

3. Storage: Elasticsearch

Elasticsearch is the "Search Engine" of the logging world.

• Time-based Indexing: Create a new index every day (e.g., logs-2024-04-20). This makes deleting old data as simple as deleting an index.
• Sharding: Distribute the index across multiple nodes to handle the write volume.

4. Cost Optimization: Tiered Storage

Logs grow exponentially. Storing everything on expensive SSDs is impossible.

• Hot Tier (SSDs): Last 24-48 hours of logs. High-speed searching.
• Warm Tier (HDDs): Last 7 days of logs. Slower, but cheaper.
• Cold Tier (S3): Logs older than 7 days. Compressed and archived for compliance.

5. Avoiding the "Feedback Loop"

The logging system should never log its own logs to the same pipeline. If a logging error occurs, it could create an infinite loop that crashes the entire infrastructure.

Summary

Building a logging system at scale is a Data Engineering challenge. By using Kafka as a buffer and Elasticsearch with time-based indexing and tiered storage, you can build a platform that provides deep visibility into your systems without breaking the bank.

Engineering Standard: The "Staff" Perspective

In high-throughput distributed systems, the code we write is often the easiest part. The difficulty lies in how that code interacts with other components in the stack.

1. Data Integrity and The "P" in CAP

Whenever you are dealing with state (Databases, Caches, or In-memory stores), you must account for Network Partitions. In a standard Java microservice, we often choose Availability (AP) by using Eventual Consistency patterns. However, for financial ledgers, we must enforce Strong Consistency (CP), which usually involves distributed locks (Redis Redlock or Zookeeper) or a strictly linearizable sequence.

2. The Observability Pillar

Writing logic without observability is like flying a plane without a dashboard. Every production service must implement:

• Tracing (OpenTelemetry): Track a single request across 50 microservices.
• Metrics (Prometheus): Monitor Heap usage, Thread saturation, and P99 latencies.
• Structured Logging (ELK/Splunk): Never log raw strings; use JSON so you can query logs like a database.

3. Production Incident Prevention

To survive a 3:00 AM incident, we use:

• Circuit Breakers: Stop the bleeding if a downstream service is down.
• Bulkheads: Isolate thread pools so one failing endpoint doesn't crash the entire app.
• Retries with Exponential Backoff: Avoid the "Thundering Herd" problem when a service comes back online.

Critical Interview Nuance

When an interviewer asks you about this topic, don't just explain the code. Explain the Trade-offs. A Staff Engineer is someone who knows that every architectural decision is a choice between two "bad" outcomes. You are picking the one that aligns with the business goal.

Performance Checklist for High-Load Systems:

1. Minimize Object Creation: Use primitive arrays and reusable buffers.
2. Batching: Group 1,000 small writes into 1 large batch to save I/O cycles.
3. Async Processing: If the user doesn't need the result immediately, move it to a Message Queue (Kafka/SQS).

Advanced Architectural Blueprint: The Staff Perspective

In modern high-scale engineering, the primary differentiator between a Senior and a Staff Engineer is the ability to see beyond the local code and understand the Global System Impact. This section provides the exhaustive architectural context required to operate this component at a "MANG" (Meta, Amazon, Netflix, Google) scale.

1. High-Availability and Disaster Recovery (DR)

Every component in a production system must be designed for failure. If this component resides in a single availability zone, it is a liability.

• Multi-Region Active-Active: To achieve "Five Nines" (99.999%) availability, we replicate state across geographical regions using asynchronous replication or global consensus (Paxos/Raft).
• Chaos Engineering: We regularly inject "latency spikes" and "node kills" using tools like Chaos Mesh to ensure the system gracefully degrades without a total outage.

2. The Data Integrity Pillar (Consistency Models)

When managing state, we must choose our position on the CAP theorem spectrum.

3. Observability and "Day 2" Operations

Writing the code is only 10% of the lifecycle. The remaining 90% is spent monitoring and maintaining it.

• Tracing (OpenTelemetry): We use distributed tracing to map the request flow. This is critical when a P99 latency spike occurs in a mesh of 100+ microservices.
• Structured Logging: We avoid unstructured text. Every log line is a JSON object containing correlationId, tenantId, and latencyMs.
• Custom Metrics: We export business-level metrics (e.g., "Orders processed per second") to Prometheus to set up intelligent alerting with PagerDuty.

4. Production Readiness Checklist for Staff Engineers

• Capacity Planning: Have we performed load testing to find the "Breaking Point" of the service?
• Security Hardening: Is all communication encrypted using mTLS (Mutual TLS)?
• Backpressure Propagation: Does the service correctly return HTTP 429 or 503 when its internal thread pools are saturated?
• Idempotency: Can the same request be retried 10 times without side effects? (Critical for Payment systems).

Critical Interview Reflection

When an interviewer asks "How would you improve this?", they are looking for your ability to identify Bottlenecks. Focus on the network I/O, the database locking strategy, or the memory allocation patterns of the JVM. Explain the trade-offs between "Throughput" and "Latency." A Staff Engineer knows that you can never have both at their theoretical maximums.

Optimization Summary:

1. Reduce Context Switching: Use non-blocking I/O (Netty/Project Loom).
2. Minimize GC Pressure: Prefer primitive specialized collections over standard Generics.
3. Data Sharding: Use Consistent Hashing to avoid "Hot Shards."

Technical Trade-offs: Messaging Systems

Key Takeaways

• High Throughput: Ingesting millions of log lines per second.
• Searchability: Full-text search across logs (errors, request IDs).
• Retention: Keeping "hot" logs for 7 days and "cold" logs for 1 year.

Read Next

• Complete System Design Interview Preparation Roadmap
• Beyond CAP: The PACELC Theorem for Distributed Databases
• Graceful Degradation: Feature Shedding

Verbal Interview Script

Interviewer: "How would you ensure high availability and fault tolerance for this specific architecture?"

Candidate: "To achieve 'Five Nines' (99.999%) availability, we must eliminate all Single Points of Failure (SPOF). I would deploy the API Gateway and stateless microservices across multiple Availability Zones (AZs) behind an active-active load balancer. For the data layer, I would use asynchronous replication to a read-replica in a different region for disaster recovery. Furthermore, it's not enough to just deploy redundantly; we must protect the system from cascading failures. I would implement strict timeouts, retry mechanisms with exponential backoff and jitter, and Circuit Breakers (using a library like Resilience4j) on all synchronous network calls between microservices."