looking for help with ElasticCache We use ElasticCache Redis to run Resque-based Qing. this means it's a mixture of sorted sets and lists. during normal operation, everything is in order, and we see good response time and throughput. The processor level is about 7-10%, the Get + Set commands are about 120-140 thousand operations. (All metrics are based on events in seconds). but - when the system experiences a (soft) burst of data by inserting a few K messages, we see that the server becomes almost unresponsive. The CPU is resistant to 100% use (the metric says 50, but uses one core) the number of operations drops to ~ 10K, the response time does not depend on SECONDS per request We expect that even if the processor is loaded to such an extent, the throughput level would remain the same. this is what we experience when running Redis locally. redis may use a processor, but bandwidth remains high. since it is initially single-line, not context switching. AFAWK - we do NOT impose any restrictions or perseverance, without replication. using the basic configuration.
size: cache.r3.large we do not use a periodic snapshot
source share