Türker Yıldırım

redis

  • Seven services in one development environment

    Deep dive10 min read

    Apache, Nginx, PHP, MySQL, Redis, Beanstalkd and a mail catcher, on a laptop. Onboarding measured in minutes rather than days, and what it costs to run.

  • Consumer groups, and the pending entries list

    Note2 min read

    Reading a stream directly gives every consumer every message. A consumer group distributes entries between members and remembers which ones were handed out and not acknowledged. The pending…

  • Cache the computation, not the query result

    Note2 min read

    Caching the rows a query returned saves the database round trip and leaves the hydration, the mapping and the aggregation to happen on every request anyway. Profiling usually…

  • A cache stampede, and the lock that prevents it

    Note2 min read

    An expensive cached value expires under load and every concurrent request misses simultaneously, so forty requests all run the same four-second query at once and the database falls…

  • A queue table is not a queue

    Deep dive10 min read

    SELECT … WHERE processed = 0 LIMIT 1, and two workers taking the same row. SKIP LOCKED makes the table viable; it does not make it a broker.

  • A stream is not a list, and XADD is not RPUSH

    Note1 min read

    A list used as a queue loses the message when a consumer crashes mid-work, because BRPOP removes it and there is no record that anything was in flight.…

  • XAUTOCLAIM does not exist yet, so XCLAIM and a timer

    Note1 min read

    When a consumer dies its pending entries belong to a name that will never acknowledge them, so something has to reassign them — and in Redis 5 that…

  • A Redis container without persistence is a deliberate choice

    Note2 min read

    The official Redis image saves to disk by default, and a cache that survives a restart is occasionally what you want and usually a way to carry a…

  • Redis modules exist now, and mostly you do not need one

    Note1 min read

    4.0 added a C module API, and the ecosystem that follows — search, JSON, time series, filters — is genuinely capable. It also means running a Redis that…

  • MEMORY USAGE tells you which key is the problem

    Note1 min read

    –bigkeys reports the largest key by element count, which is not the same as the largest by memory — a hash with 500 short fields is smaller than…