Scalability, Availability & Reliability

What you'll learn

• Define scalability, availability and reliability precisely and express goals as SLIs, SLOs and error budgets.
• Reason about capacity with Little's law, Amdahl's law, queueing and back-of-the-envelope estimates.
• Scale compute and data with stateless services, load balancing, autoscaling, replication and partitioning.
• Design for availability with redundancy, failover, multi-zone and multi-region setups and DR strategies.
• Prevent and contain failures with timeouts, retries, circuit breakers, load shedding and safe deployments.
• Operate reliably with golden signals, burn-rate alerts, incident response, load testing and chaos engineering.

Syllabus

Definitions, Targets and the Maths Behind Them

1. Scalability, Availability and Reliability
2. SLIs, SLOs, SLAs and Error Budgets
3. Little's Law, Amdahl's Law and Queueing

Scaling Compute

1. Vertical vs Horizontal Scaling and Statelessness
2. Load Balancing
3. Autoscaling

Scaling Data

1. Replication
2. Partitioning and Sharding
3. Consistency Trade-offs: CAP, PACELC and Quorums

Performance Under Load

1. Caching as a Scaling Tool
2. Queues, Load Levelling and Back-Pressure
3. Tail Latency and Fan-Out

Designing for Availability

1. Redundancy and Failover
2. Zones, Regions and Disaster Recovery
3. Availability Arithmetic

Failure Modes and Defences

1. How Distributed Systems Fail
2. Timeouts, Retries, Circuit Breakers and Load Shedding
3. Safe Changes: Deployments, Flags and Rollbacks

Observability, Incidents and Testing

1. Golden Signals and SLO-Based Alerting
2. Incident Response and Postmortems
3. Load Testing, Capacity Planning and Chaos Engineering

Putting It Into Practice and Revision

1. Back-of-the-Envelope Capacity Estimation
2. Case Study: Scaling a Web App Step by Step
3. Production Readiness Review
4. Revision and Interview Questions