How database indexing works, with Postgres examples
B-trees in plain English, reading EXPLAIN ANALYZE, composite, partial and covering indexes, what indexes cost, and adding them to a live Postgres table safely.
Engineering at Scale
Building for 100,000 students at once
37 articles
Articles
Read replicas and database replication explained
How PostgreSQL streaming replication works, synchronous vs asynchronous commits, routing reads in Django, handling replication lag, and replicas for failover.
Video transcoding pipelines: from upload to HLS
How to turn an uploaded lecture into encrypted HLS: probing, encoding a ladder with FFmpeg, queue-driven workers, CPU vs GPU vs spot capacity, retries and what one lecture really costs.
Adaptive bitrate streaming and bitrate ladders explained
How players choose between renditions, what hls.js and dash.js do by default, and how to design a bitrate ladder that keeps lectures playing on patchy mobile networks.
HLS vs DASH: how the formats compare, and where CMAF fits
HLS and MPEG-DASH compared: playlists vs manifests, iPhone and Android support, encryption and DRM, low latency, and how CMAF lets one set of segments serve both formats.
WebRTC vs HLS for live classes: latency, scale and cost
Sub-second WebRTC or CDN-scale HLS? Latency budgets for live classes, SFUs, Low-Latency HLS, RTMP vs SRT ingest, the cost of a 5,000-student class and a hybrid design.
Multi-CDN and CDN failover for video platforms
Why one CDN is a single point of failure on exam and launch days, how DNS, player-side and content-steering failover compare, and what must match across CDNs.
Message queues and background jobs: Celery, Redis, Kafka
Move slow work off the request path: queues vs brokers vs streams, Celery in Django, retries and idempotency, dead letters, separate queues, and when Kafka is worth it.
Monitoring vs observability: logs, metrics, traces and SLOs
What monitoring and observability each answer, how logs, metrics and traces fit together with OpenTelemetry, and how to set SLOs and error budgets around student journeys.
RTO and RPO: planning disaster recovery that works
Set recovery time and recovery point objectives per system, choose between backup and restore, pilot light, warm standby and active/active, apply 3-2-1 backups and run drills.
Zero-downtime deployment: blue-green, canary and rolling
Deploy during a live class without anyone noticing: rolling, blue-green and canary compared, draining connections and WebSockets, feature flags and exam-day deploy freezes.
Zero-downtime database migrations: expand and contract
Change a live Postgres schema without locking students out: expand and contract, lock timeouts, concurrent indexes, NOT VALID constraints, batched backfills, safe Django migrations.
Upclass for institutes
Launch your academy. Keep every course safe.
Your courses, live classes and tests in your own branded app, with piracy protection on every lecture. Plans start at ₹1,999/month.