Redis / Cache

Why individually atomic Redis commands do not make multi-command application workflows atomic, and how INCR, WATCH, Lua/Functions, locks, and database constraints solve different races.

What Redis Slow Log measures, what it omits, how to configure and inspect it, and how to correlate slow commands with end-to-end latency.

A production-oriented guide to Redis stalls caused by slow commands, big keys, CPU saturation, persistence/fork, memory pressure, swap, networking, and client backpressure.

Why oversized Redis values or collections cause latency, bandwidth, memory, replication, deletion, and cluster-balance problems, plus safe detection and mitigation.

How one disproportionately popular key overloads a Redis shard, how to detect it, and strategies such as local caches, replica reads, key replication, request coalescing, and data-model changes.

Redis Cluster sharding with 16,384 hash slots, client redirection, hash tags, resharding, replica failover, and its consistency and multi-key constraints.

Redis Sentinel monitoring and failover for primary-replica deployments, including subjective/objective down states, quorum, failover authorization, promotion, and remaining consistency limits.

Redis primary-replica replication, full and partial synchronization, replication backlog, stale replica reads, and asynchronous failure semantics.

TTL semantics and Redis's combined passive and active expiration strategy, without depending on historical sampling constants.

How Redis memory is consumed, why RSS differs from logical dataset memory, fragmentation and allocator effects, maxmemory, expiry, eviction, and practical observability.