Sutty/sutty
5
0
Fork 0
Code Issues 2 Pull requests Projects Releases Wiki Activity

recolección de estadísticas con mejor performance

Browse source
This commit is contained in:
f 2021-10-20 13:14:19 -03:00
parent 9cf7c62861
commit 8d38d0d2ae
1 changed files with 26 additions and 16 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

30
app/jobs/stat_collection_job.rb
View file

@ -21,14 +21,20 @@ class StatCollectionJob < ApplicationJob
COLUMNS = %i[uri].freeze COLUMNS = %i[uri].freeze
def perform(once: false) def perform(once: false)
Stat::INTERVALS.each do |interval| Site.find_each do |site|
options = { interval: interval } hostnames = [site.hostname, site.alternative_hostnames].flatten
# Usamos el primero porque luego podemos hacer un rollup recursivo
options = { interval: Stat::INTERVALS.first }
# Visitas por hostname # Visitas por hostname
AccessLog.completed_requests.non_robots.pages.group(:host).rollup('host', **options) hostnames.each do |hostname|
AccessLog.where(host: hostname).completed_requests.non_robots.pages.group(:host).rollup('host', **options)
combined_columns(**options) combined_columns(hostname, **options)
stats_by_site(**options) end
stats_by_site(site, **options)
end end
# Registrar que se hicieron todas las recolecciones # Registrar que se hicieron todas las recolecciones
@ -40,9 +46,15 @@ class StatCollectionJob < ApplicationJob
private private
# Combinación de columnas # Combinación de columnas
def combined_columns(**options) def combined_columns(hostname, **options)
where = { host: hostname }
COLUMNS.each do |column| COLUMNS.each do |column|
AccessLog.completed_requests.non_robots.group(:host, column).rollup("host|#{column}", **options) AccessLog.where(host: hostname).pluck(Arel.sql("distinct #{column}")).each do |value|
where[column] = value
AccessLog.where(**where).completed_requests.non_robots.group(:host, column).rollup("host|#{column}", **options)
end
end end
end end
@ -50,8 +62,7 @@ class StatCollectionJob < ApplicationJob
# #
# XXX: En realidad se agrupan por el deploy_id, que siempre será el # XXX: En realidad se agrupan por el deploy_id, que siempre será el
# del DeployLocal. # del DeployLocal.
def stats_by_site(**options) def stats_by_site(site, **options)
Site.find_each do |site|
site.build_stats.jekyll.group(:deploy_id).rollup('builds', **options) site.build_stats.jekyll.group(:deploy_id).rollup('builds', **options)
site.build_stats.jekyll.group(:deploy_id).rollup('space_used', **options) do |rollup| site.build_stats.jekyll.group(:deploy_id).rollup('space_used', **options) do |rollup|
@ -62,5 +73,4 @@ class StatCollectionJob < ApplicationJob
rollup.average(:seconds) rollup.average(:seconds)
end end
end end
end
end end