kbrock

2 renderings of a vmware dc folder hierarchy

Full diagram

[](https://mermaid-js.github.io/mermaid-live-editor/#/edit/eyJjb2RlIjoiZmxvd2NoYXJ0IExSXG5cbmRmW1wiREMgRiAoWSlcIl1cbmRjW1wiREMgKFkpXCJdXG5oZltcIkggRiAoWSlcIl1cbnZmW1wiVm0gRiAoQilcIl1cbmNbXCJDbHVzdGVyXCJdXG5oW1wiSG9zdFwiXVxudm0oXCJWbVwiKVxuc

kbrock

relationships = Relationship.where(relationship: 'genealogy', resource_type: 'VmOrTemplate').select(:id, :resource_id, :ancestry).to_a
rel_by_rel_id = relationships.index_by(:id) # relationship => resource_id
rel_by_vm_id = relationships.index_by(:resource_id) # resource_id => ancestry

Vm.where(:id => Relationship.select(:resource_id)).each_in_batches do |vm|
  ancestry = rel_by_vm_id[vm.id].ancestry.split("/").map { |rel_id| rel_by_rel_id[rel_id].resource_id.to_s }.join("/")
  vm.update_attribute(:ancestry => ancestry)
end

kbrock

curl -X POST http://admin:smartvm@localhost:3000/api/service_catalogs/#{id}/service_templates/1 -d '{"action" : "order"}'

kbrock

The [original] is modified in the following ways:

• tweaked options to be {} not [] for json syntax reasons
• added "R" for demonstrative purposes

Here are a few ideas. I tried to keep them as independent as possible:

2. assume width, height, matrix, and display
3. remove replaces 4. condense options want to keep nodes separate.
4. assume, replaces, condense

kbrock

#!/usr/bin/env ruby

require "benchmark/ips"
require "date"

BEGIN_OF_JULY = Date.new(2015, 7, 1)
END_OF_JULY = Date.new(2015, 7, 31)
DAY_IN_JULY = Date.new(2015, 7, 15)
RANGE = BEGIN_OF_JULY..END_OF_JULY

kbrock

method	nil	speed	str	speed
?split:[]	9,761,148.9 i/s		2,869,972.9 i/s	3.40x
&.split||[]	9,271,343.0 i/s	same	2,879,248.1 i/s	3.39x
&&split||[]	9,096,687.7 i/s	same	2,941,303.8 i/s	3.32x
to_s.split	4,288,858.0 i/s	2.28x	2,822,293.7 i/s	3.46x

ruby -v

kbrock

ruby 2.4.2p198 (2017-09-14 revision 59899) [x86_64-darwin16]

	nil	slower	empty	slower	full	slower
x&.empty?	11985357.7 i/s	same	11617138.6 i/s	same	11247144.5 i/s	same
x.blank?	9909235.8 i/s	1.21x	10488507.9 i/s	1.14x	10418916.8 i/s	1.15x
x.nil? || x.empty?	10194622.1 i/s	1.18x	9584369.0 i/s	1.25x	9667811.7 i/s	1.24x
x.try!(:empty?)	5437472.5 i/s	2.20x	3493090.9 i/s	3.43x	3523153.2 i/s	3.40x

kbrock

note: in the tests, we are setting a different variable @x than the guard clause @x2. I tested a terciary ?: vs the guard clause return if defined?() and they were about the same for pass and fail case

test	int i/s	slower	text i/s	slower
eq2	6,315,053.7 i/s	same	6,323,409.7 i/s	base
neq2	6,281,137.2 i/s	same	3,707,761.0 i/s	1.71x
eq	5,978,924.2 i/s	same	4,572,438.8 i/s	1.38x
include	4,368,299.8 i/s	1.45x	4,349,531.4 i/s	1.45x
neq	4,194,756.5 i/s	1.51x	4,170,358.7 i/s	1.52x

kbrock

Looking at the difference between using ActiveSupport::Notification, and rolling our own metrics counters.

When logging, Notifications are 66% slower than manually adding the metrics code. The metrics code is actually 50% slower when not logging the results (no subscriber). Having no subscriber does speed up performance for all the cases.

The difference for timing for payload is probably due to allocations. Having a fixed payload ended up being equivalent to the no payload case. This is a reminder for us to pay attention to what we pass in the payload.

kbrock

gem
Json oj	40.4 i/s
MessagePack	38.3 i/s	same
binary json	22.0 i/s	1.84x slower
Json yajl	19.0 i/s	2.13x slower
ruby json	15.6 i/s	2.59x slower
ruby marshal	13.9 i/s	2.90x slower
ruby yaml	0.4 i/s	111.99x slower
Yaml Psych	0.4 i/s	112.09x slower