Connected clients:0 (Garnet 1.0.15)

[Andrey Kuvshinov (everanium)]

Maybe I don’t understand something, but why does it show that 0 clients are connected?

root@srv:/var/log/garnet# redis-cli -p 7333
127.0.0.1:7333> info clients
# Clients
connected_clients:0
127.0.0.1:7333> exit
root@srv:/var/log/garnet# netstat -tapn | grep 7333
tcp        0      0 127.0.0.1:7333          0.0.0.0:*               LISTEN      2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53026         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53026         127.0.0.1:7333          ESTABLISHED 2419840/angie: work
tcp        0      0 127.0.0.1:53126         127.0.0.1:7333          ESTABLISHED 2419841/angie: work
tcp        0      0 127.0.0.1:53110         127.0.0.1:7333          ESTABLISHED 2419831/angie: work
tcp        0      0 127.0.0.1:53130         127.0.0.1:7333          ESTABLISHED 2419946/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53130         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53124         127.0.0.1:7333          ESTABLISHED 2419833/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:40200         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53084         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53142         127.0.0.1:7333          ESTABLISHED 2419832/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53158         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:40200         127.0.0.1:7333          ESTABLISHED 2420393/redis-cli
tcp        0      0 127.0.0.1:7333          127.0.0.1:53040         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53158         127.0.0.1:7333          ESTABLISHED 2419867/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53068         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53072         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53166         127.0.0.1:7333          ESTABLISHED 2419834/angie: work
tcp        0      0 127.0.0.1:53054         127.0.0.1:7333          ESTABLISHED 2419836/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53124         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:47062         127.0.0.1:7333          TIME_WAIT   -
tcp        0      0 127.0.0.1:7333          127.0.0.1:53110         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53052         127.0.0.1:7333          ESTABLISHED 2419843/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53166         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53054         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53100         127.0.0.1:7333          ESTABLISHED 2419844/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53052         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53072         127.0.0.1:7333          ESTABLISHED 2419839/angie: work
tcp        0      0 127.0.0.1:53104         127.0.0.1:7333          ESTABLISHED 2419842/angie: work
tcp        0      0 127.0.0.1:53068         127.0.0.1:7333          ESTABLISHED 2419838/angie: work
tcp        0      0 127.0.0.1:53084         127.0.0.1:7333          ESTABLISHED 2419835/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53142         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:53040         127.0.0.1:7333          ESTABLISHED 2419837/angie: work
tcp        0      0 127.0.0.1:7333          127.0.0.1:53126         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53104         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:7333          127.0.0.1:53100         ESTABLISHED 2419038/GarnetServe
tcp        0      0 127.0.0.1:36170         127.0.0.1:7333          TIME_WAIT   -

[Vasileios Zois (vazois)]

I know it is confusing, but server metrics are not turned on by default. You have to use --metrics-sampling-freq flag and --latency-monitor if you want to track latency histograms.
We did this to avoid perf impact in the critical path. It is likely that info type metrics do not impact performance that much, though we have to evaluate it thoroughly.

[Andrey Kuvshinov (everanium)]

Thanks