$ syswork --explain
See how systems actually work. Then break them.
Animated simulations of real system internals: message brokers, load balancers, replication, transactions. Stall a consumer, cut a network link, crash a node, or pull a component out entirely. Then watch, step by step, what the system does and why.
Pub/Sub
liveFire-and-forget fan-out. Stall a subscriber, cut its network or crash Redis, and see where messages go: client output buffers, soft and hard limits, half-open connections.
Streams & consumer groups
liveAn append-only log with consumer groups. Kill, hang or poison a consumer and watch the pending entries list (PEL), XACK, XAUTOCLAIM, dead-lettering and MAXLEN trimming.
Produce & replicate
plannedPartitioner, leader and followers pulling (fetching) from the leader, ISR, high watermark, acks=0/1/all, min.insync.replicas, leader failover.
Consumer groups
plannedPull-based consumers, lag, offset commits and rebalancing.
Algorithms side by side
plannedRound robin, weighted, least connections, power of two choices, under uneven servers and slow requests.
Consistent hashing
plannedThe hash ring, virtual nodes, and how many keys move when a node joins or leaves.
Raft leader election
plannedTerms, randomized election timeouts, votes, split votes, log replication and commit index under partitions.
MongoDB replica set
plannedPrimary election, oplog replication, write concern majority, and rollback after a failover.
Isolation levels & anomalies
plannedTwo transactions on parallel timelines: dirty reads, non-repeatable reads, lost updates and write skew at each isolation level.
WAL & crash recovery
plannedWhy a committed transaction survives a crash: write-ahead log, fsync, checkpoints and replay.