|
#!/usr/bin/env ruby
require 'cgi'
require 'httparty'
require 'date'
module Agents
class JiraAgent < Agent
cannot_receive_events!
description <<-MD
The Jira Agent subscribes to Jira issue updates.
- `jira_url` specifies the full URL of the jira installation, including https://
- `jql` is an optional Jira Query Language-based filter to limit the flow of events. See [JQL Docs](https://confluence.atlassian.com/display/JIRA/Advanced+Searching) for details.
- `username` and `password` are optional, and may need to be specified if your Jira instance is read-protected
- `timeout` is an optional parameter that specifies how long the request processing may take in minutes.
The agent does periodic queries and emits the events containing the updated issues in JSON format.
NOTE: upon the first execution, the agent will fetch everything available by the JQL query. So if it's not desirable, limit the `jql` query by date.
MD
event_description <<-MD
Events are the raw JSON generated by Jira REST API
{
"expand": "editmeta,renderedFields,transitions,changelog,operations",
"id": "80127",
"self": "https://jira.atlassian.com/rest/api/2/issue/80127",
"key": "BAM-3512",
"fields": {
...
}
}
MD
default_schedule "every_10m"
MAX_EMPTY_REQUESTS = 10
def default_options
{
'username' => '',
'password' => '',
'jira_url' => 'https://jira.atlassian.com',
'jql' => '',
'expected_update_period_in_days' => '7',
'timeout' => '1'
}
end
def validate_options
errors.add(:base, "you need to specify password if user name is set") if options['username'].present? and not options['password'].present?
errors.add(:base, "you need to specify your jira URL") unless options['jira_url'].present?
errors.add(:base, "you need to specify the expected update period") unless options['expected_update_period_in_days'].present?
errors.add(:base, "you need to specify request timeout") unless options['timeout'].present?
end
def working?
event_created_within?(interpolated['expected_update_period_in_days']) && !recent_error_logs?
end
def check
last_run = nil
current_run = Time.now.utc.iso8601
last_run = Time.parse(memory[:last_run]) if memory[:last_run]
issues = get_issues(last_run)
issues.each do |issue|
updated = Time.parse(issue['fields']['updated'])
# this check is more precise than in get_issues()
# see get_issues() for explanation
if not last_run or updated > last_run
create_event :payload => issue
end
end
memory[:last_run] = current_run
end
private
def request_url(jql, start_at)
"#{interpolated[:jira_url]}/rest/api/2/search?jql=#{CGI::escape(jql)}&fields=*all&startAt=#{start_at}"
end
def request_options
ropts = {:headers => {"User-Agent" => "Huginn (https://github.com/cantino/huginn)"}}
if !interpolated[:username].empty?
ropts = ropts.merge({:basic_auth => {:username =>interpolated[:username], :password=>interpolated[:password]}})
end
ropts
end
def get(url, options)
response = HTTParty.get(url, options)
if response.code == 400
raise RuntimeError.new("Jira error: #{response['errorMessages']}")
elsif response.code == 403
raise RuntimeError.new("Authentication failed: Forbidden (403)")
elsif response.code != 200
raise RuntimeError.new("Request failed: #{response}")
end
response
end
def get_issues(since)
startAt = 0
issues = []
# JQL doesn't have an ability to specify timezones
# Because of this we have to fetch issues 24 h
# earlier and filter out unnecessary ones at a later
# stage. Fortunately, the 'updated' field has GMT
# offset
since -= 24*60*60 if since
jql = ""
if !interpolated[:jql].empty? && since
jql = "(#{interpolated[:jql]}) and updated >= '#{since.strftime('%Y-%m-%d %H:%M')}'"
else
jql = interpolated[:jql] if !interpolated[:jql].empty?
jql = "updated >= '#{since.strftime('%Y-%m-%d %H:%M')}'" if since
end
start_time = Time.now
request_limit = 0
loop do
response = get(request_url(jql, startAt), request_options)
if response['issues'].length == 0
request_limit+=1
end
if request_limit > MAX_EMPTY_REQUESTS
raise RuntimeError.new("There is no progress while fetching issues")
end
if Time.now > start_time + interpolated['timeout'].to_i * 60
raise RuntimeError.new("Timeout exceeded while fetching issues")
end
issues += response['issues']
startAt += response['issues'].length
break if startAt >= response['total']
end
issues
end
end
end
|