Could you plz add maxclients param into info clients? It will be useful for monitoring like prometheus or telegraf.

Comment From: oranagra

good idea https://github.com/redis/redis/pull/7979

Comment From: ShooterIT

How about showing remaining available clients number @iglov @oranagra

Comment From: iglov

Hey guys! Thanks 4 ur super fast reaction, i appreciate that! :) @ShooterIT i dunno, ok if you want. I planned to do that with my prometheus: maxclients - connected_clients - that the reason of this FR and maxclients param :)

Comment From: oranagra

@ShooterIT in some cases INFO is indeed designed to be console viewer friendly (like showing used memory in human units, and showing the instantaneous metrics), but in many places it's more suitable to be processed by software.

I think in this case the metric is more intended to be processed by software, so i think it's better as maxclients, also i'm trying to follow the existing exposed maxmemory which is also why i used maxclients (same exact name as the CONFIG setting) and not max_clients or such.

there's at least one place i'm aware of that took the other approach, and i think it was a mistake: showing master_sync_left_bytes instead of repl_transfer_size and repl_transfer_read. unlike the approach in showing loading_total_bytes and loading_loaded_bytes.

maybe i should use this opportunity to expose repl_transfer_size and / or repl_transfer_read?

Comment From: ShooterIT

@iglov I guess :), I ever wanted to do this, but actually if you want to get remaining available clients number when enable cluster-mode, you should minus cluster bus link number that doesn't seem to show.

@oranagra the existing exposed maxmemory convince me, i feel OK.

Comment From: oranagra

have a look at this: https://github.com/redis/redis/pull/7981

Comment From: ShooterIT

Hi @iglov FYI https://github.com/redis/redis-doc/pull/1427/files#diff-a41c08daecd2572f692ee897da8837a4da089dab5291b6ccfd6d63e9450231fcR78-R80

Comment From: iglov

@ShooterIT hey there! Yeah, i've already read it, LG2M, thx!