fix(ioredis): set net.peer.name attribute according to spec

[blumamir]

Which problem is this PR solving?

fixes #240

Short description of the changes

replace attribute net.peer.host with net.peer.name so it follows the specification Semantic conventions for database client calls

[codecov]

Codecov Report

Merging #241 (f9e86ae) into master (12749ae) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #241   +/-   ##
=======================================
  Coverage   94.95%   94.95%           
=======================================
  Files         115      115           
  Lines        5990     5990           
  Branches      618      618           
=======================================
  Hits         5688     5688           
  Misses        302      302           

[naseemkullah]

Should we parse the host to see if it is an IP and if so set NET_PEER_IP instead?

[blumamir]

Should we parse the host to see if it is an IP and if so set NET_PEER_IP instead?

In my opinion it makes sense this way, because we populate the attribute from what the user configured, and not extracting IP address from the socket itself like the HTTP plugin is doing.
But I can see how it also makes sense to use the NET_PEER_IP.
If you think NET_PEER_IP is preferable I'll change it.

[naseemkullah]

What I meant is that a host set by the user can either be a name or an IP. If we choose one PEER NAME is good, but ideally we check if the host is an IP and if so we set pEER IP but if it's a name we set PEER NAME

[naseemkullah]

If you do choose to implement this check the following SO post might help. If not it's fine this lgtm since it's already an improvement
https://stackoverflow.com/questions/4460586/javascript-regular-expression-to-check-for-ip-addresses

[naseemkullah]

In reality the spec should probably just have PEER HOST