Rad s RabbitMQ-om zahtijeva praćenje ključnih metrika (dubina reda čekanja, brzine poruka, zdravlje potrošača, resursi) i korištenje alata za upravljanje. Razumijevanje praćenja i upravljanja važno je za pouzdanu konfiguraciju RabbitMQ-a.
Ključne metrike za praćenje
✓ QUEUE DEPTH (length) → growing queues = consumers can't keep up (a key signal!) — like
consumer lag; investigate (add consumers, fix slow processing)
✓ MESSAGE RATES → publish rate vs deliver/ack rate (in vs out — are they balanced?)
✓ CONSUMER count and health → are consumers connected and processing?
✓ UNACKED messages → many unacked = slow/stuck consumers
✓ RESOURCES → memory, disk, CPU, connections, file descriptors (RabbitMQ has memory/disk
alarms that block publishing when thresholds are hit!)
✓ DEAD LETTER queue size → failed messages accumulating (signals problems)
