2
0
Fork 0
mirror of https://github.com/discourse/discourse.git synced 2025-10-03 17:21:20 +08:00

DEV: Add DB backed problem checks to support perform_every config (#25834)

As part of problem checks refactoring, we're moving some data to be DB backed. In this PR it's the tracking of problem check execution. When was it last run, when was the last problem, when should it run next, how many consecutive checks had problems, etc.

This allows us to implement the perform_every feature in scheduled problem checks for checks that don't need to be run every 10 minutes.
This commit is contained in:
Ted Johansson 2024-02-27 11:17:39 +08:00 committed by GitHub
parent 412b36cc93
commit 1bcb521fbf
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
8 changed files with 230 additions and 15 deletions

View file

@ -18,7 +18,12 @@ module Jobs
problems = check.call
raise RetrySignal if problems.present? && retry_count < check.max_retries

problems.each { |problem| AdminDashboardData.add_found_scheduled_check_problem(problem) }
if problems.present?
problems.each { |problem| AdminDashboardData.add_found_scheduled_check_problem(problem) }
ProblemCheckTracker[identifier].problem!(next_run_at: check.perform_every.from_now)
else
ProblemCheckTracker[identifier].no_problem!(next_run_at: check.perform_every.from_now)
end
rescue RetrySignal
Jobs.enqueue_in(
check.retry_after,

View file

@ -14,7 +14,12 @@ module Jobs
# not be re-added by the relevant checker, and will be cleared.
AdminDashboardData.clear_found_scheduled_check_problems

::ProblemCheck.scheduled.each do |check|
scheduled_checks =
ProblemCheckTracker.all.filter_map do |tracker|
tracker.check if tracker.check.scheduled? && tracker.ready_to_run?
end

scheduled_checks.each do |check|
Jobs.enqueue(:problem_check, check_identifier: check.identifier.to_s)
end
end

View file

@ -0,0 +1,47 @@
# frozen_string_literal: true

class ProblemCheckTracker < ActiveRecord::Base
validates :identifier, presence: true, uniqueness: true
validates :blips, presence: true, numericality: { greater_than_or_equal_to: 0 }

def self.[](identifier)
find_or_create_by(identifier:)
end

def ready_to_run?
next_run_at.blank? || next_run_at.past?
end

def problem!(next_run_at: nil)
now = Time.current

update!(blips: blips + 1, last_run_at: now, last_problem_at: now, next_run_at:)
end

def no_problem!(next_run_at: nil)
now = Time.current

update!(blips: 0, last_run_at: now, last_success_at: now, next_run_at:)
end

def check
ProblemCheck[identifier]
end
end

# == Schema Information
#
# Table name: problem_check_trackers
#
# id :bigint not null, primary key
# identifier :string not null
# blips :integer default(0), not null
# last_run_at :datetime
# next_run_at :datetime
# last_success_at :datetime
# last_problem_at :datetime
#
# Indexes
#
# index_problem_check_trackers_on_identifier (identifier) UNIQUE
#

View file

@ -0,0 +1,16 @@
# frozen_string_literal: true

class CreateProblemCheckTrackers < ActiveRecord::Migration[7.0]
def change
create_table :problem_check_trackers do |t|
t.string :identifier, null: false, index: { unique: true }

t.integer :blips, null: false, default: 0

t.datetime :last_run_at
t.datetime :next_run_at
t.datetime :last_success_at
t.datetime :last_problem_at
end
end
end

View file

@ -0,0 +1,3 @@
# frozen_string_literal: true

Fabricator(:problem_check_tracker) {}

View file

@ -11,6 +11,7 @@ RSpec.describe Jobs::ProblemCheck do
before do
::ProblemCheck::TestCheck =
Class.new(::ProblemCheck) do
self.perform_every = 30.minutes
self.max_retries = 0

def call
@ -39,6 +40,7 @@ RSpec.describe Jobs::ProblemCheck do
before do
::ProblemCheck::TestCheck =
Class.new(::ProblemCheck) do
self.perform_every = 30.minutes
self.max_retries = 0

def call
@ -71,6 +73,7 @@ RSpec.describe Jobs::ProblemCheck do
before do
::ProblemCheck::TestCheck =
Class.new(::ProblemCheck) do
self.perform_every = 30.minutes
self.max_retries = 2

def call
@ -79,6 +82,12 @@ RSpec.describe Jobs::ProblemCheck do
end
end

it "does not yet update the problem check tracker" do
expect {
described_class.new.execute(check_identifier: :test_check, retry_count: 1)
}.not_to change { ProblemCheckTracker.where("blips > ?", 0).count }
end

it "schedules a retry" do
expect_enqueued_with(
job: :problem_check,
@ -94,6 +103,7 @@ RSpec.describe Jobs::ProblemCheck do
before do
::ProblemCheck::TestCheck =
Class.new(::ProblemCheck) do
self.perform_every = 30.minutes
self.max_retries = 1

def call
@ -102,9 +112,15 @@ RSpec.describe Jobs::ProblemCheck do
end
end

it "updates the problem check tracker" do
expect {
described_class.new.execute(check_identifier: :test_check, retry_count: 1)
}.to change { ProblemCheckTracker.where("blips > ?", 0).count }.by(1)
end

it "does not schedule a retry" do
expect_not_enqueued_with(job: :problem_check) do
described_class.new.execute(check_identifier: :test_identifier, retry_count: 1)
described_class.new.execute(check_identifier: :test_check, retry_count: 1)
end
end
end

View file

@ -20,9 +20,43 @@ RSpec.describe Jobs::ProblemChecks do
ProblemCheck.send(:remove_const, "NonScheduledCheck")
end

it "schedules the individual scheduled checks" do
expect_enqueued_with(job: :problem_check, args: { check_identifier: "scheduled_check" }) do
described_class.new.execute([])
context "when the tracker determines the check is ready to run" do
before do
ProblemCheckTracker.create!(identifier: "scheduled_check", next_run_at: 5.minutes.ago)
end

it "schedules the individual scheduled checks" do
expect_enqueued_with(job: :problem_check, args: { check_identifier: "scheduled_check" }) do
described_class.new.execute([])
end
end
end

context "when the tracker determines the check shouldn't run yet" do
before do
ProblemCheckTracker.create!(identifier: "scheduled_check", next_run_at: 5.minutes.from_now)
end

it "does not schedule any check" do
expect_not_enqueued_with(
job: :problem_check,
args: {
check_identifier: "scheduled_check",
},
) { described_class.new.execute([]) }
end
end

context "when dealing with a non-scheduled check" do
before { ProblemCheckTracker.create!(identifier: "non_scheduled_check", next_run_at: nil) }

it "does not schedule any check" do
expect_not_enqueued_with(
job: :problem_check,
args: {
check_identifier: "non_scheduled_check",
},
) { described_class.new.execute([]) }
end
end

@ -34,13 +68,4 @@ RSpec.describe Jobs::ProblemChecks do
},
) { described_class.new.execute([]) }
end

it "does not schedule non-scheduled checks" do
expect_not_enqueued_with(
job: :problem_check,
args: {
check_identifier: "non_scheduled_check",
},
) { described_class.new.execute([]) }
end
end

View file

@ -0,0 +1,98 @@
# frozen_string_literal: true

RSpec.describe ProblemCheckTracker do
describe "validations" do
let(:record) { described_class.new(identifier: "twitter_login") }

it { expect(record).to validate_presence_of(:identifier) }
it { expect(record).to validate_uniqueness_of(:identifier) }

it { expect(record).to validate_numericality_of(:blips).is_greater_than_or_equal_to(0) }
end

describe ".[]" do
before { Fabricate(:problem_check_tracker, identifier: "twitter_login") }

context "when the problem check tracker already exists" do
it { expect(described_class[:twitter_login]).not_to be_new_record }
end

context "when the problem check tracker doesn't exist yet" do
it { expect(described_class[:facebook_login]).to be_previously_new_record }
end
end

describe "#ready_to_run?" do
let(:problem_tracker) { described_class.new(next_run_at:) }

context "when the next run timestamp is not set" do
let(:next_run_at) { nil }

it { expect(problem_tracker).to be_ready_to_run }
end

context "when the next run timestamp is in the past" do
let(:next_run_at) { 5.minutes.ago }

it { expect(problem_tracker).to be_ready_to_run }
end

context "when the next run timestamp is in the future" do
let(:next_run_at) { 5.minutes.from_now }

it { expect(problem_tracker).not_to be_ready_to_run }
end
end

describe "#problem!" do
let(:problem_tracker) do
Fabricate(:problem_check_tracker, identifier: "twitter_login", **original_attributes)
end

let(:original_attributes) do
{
blips: 0,
last_problem_at: 1.week.ago,
last_success_at: 24.hours.ago,
last_run_at: 24.hours.ago,
next_run_at: nil,
}
end

let(:updated_attributes) { { blips: 1 } }

it do
freeze_time

expect { problem_tracker.problem!(next_run_at: 24.hours.from_now) }.to change {
problem_tracker.attributes
}.to(hash_including(updated_attributes))
end
end

describe "#no_problem!" do
let(:problem_tracker) do
Fabricate(:problem_check_tracker, identifier: "twitter_login", **original_attributes)
end

let(:original_attributes) do
{
blips: 0,
last_problem_at: 1.week.ago,
last_success_at: Time.current,
last_run_at: 24.hours.ago,
next_run_at: nil,
}
end

let(:updated_attributes) { { blips: 0 } }

it do
freeze_time

expect { problem_tracker.no_problem!(next_run_at: 24.hours.from_now) }.to change {
problem_tracker.attributes
}.to(hash_including(updated_attributes))
end
end
end