Added note to Deployment guide about a known issue in 17.13 and older  to satisfy multiple partner requests.

When we release 17.1.4, I need to modify that note. 

[NOTE]
If your RHOSP 17.1.3 or earlier deployment includes a filter rule in nftables or iptables with a LOG action, and the kernel command line (`/proc/cmdline`) has `console=tty50`, logging actions can cause substantial latency in packet transmission. 
If your deployment has this configuration and you observe excessive latency, you can apply the workaround described in Knowledgebase solution link:https://access.redhat.com/solutions/7076184[Sometimes receiving packet(e.g. ICMP echo) has latency, around 190 ms].
To track a planned fix for this issue, see link:https://bugzilla.redhat.com/show_bug.cgi?id=2293368.

////  commented out
after 17.1.4 release, delete the previous note, uncomment the following note, and delete this line. the previous note is written for 17.1.3 and older. the following note is written for 17.1.4
[NOTE]
Starting with RHOSP 17.1.4, all kernel console logging arguments are removed because console logging can cause unacceptable latency issues in Compute workloads. 
+
If your RHOSP 17.1.3 or earlier deployment includes a filter rule in nftables or iptables with a LOG action, and the kernel command line (`/proc/cmdline`) has `console=tty50`, logging actions can cause substantial latency in packet transmission. 
If your 17.1.3 or earlier deployment has this configuration and you observe excessive latency, you can update to 17.1.4 or apply the workaround described in Knowledgebase solution link:https://access.redhat.com/solutions/7076184[Sometimes receiving packet(e.g. ICMP echo) has latency, around 190 ms].

////