rt2jira.py

Converts RT tickets to JIRA tickets

Author

Darien Kindlund ([email protected])

Description

This Python script automatically convert RT tickets to JIRA tickets, where the uniqueness of the RT ticket is based on the (requestor identity, subject) of the RT ticket, which will then get assigned a unique JIRA ticket. This script will also synchronize new comments on existing RT tickets and represent them as new JIRA comments. However, this script is a 1-way sync. It will not take new JIRA comments/tickets and represent that activity in new/existing RT tickets. This script is designed to be run on a periodic basis, in order to resync new RT tickets to JIRA tickets.

Prerequisites

pip install six
pip install jira
pip install python-rtkit
pip install titlecase

Quick Start

1. Get all prerequisite libraries installed.

2. Download and extract the rt2jira package.

3. Edit and review the config.ini file.

4. Run: python rt2jira.py config.ini

5. Rerun the script as often as you want the RT queue and the JIRA tickets synchronized

config.ini Notes

The following are comments on some of the INI settings.

NOTE: Please review all settings in the config.ini file before running the script.

[rt]

• api_url_prefix: This is the URL for the RT REST API. Traditionally, it should look something like: https://rt.server.com/REST/1.0/

• api_search_suffix: Search query to feed into the REST API to pull relevant tickets down to be ported as JIRA tickets.

  For example, if this value is set to something like: search/ticket?query=Queue+%3D+%27RT-Queue-Name%27+AND+LastUpdated+%3E+%27-5+days%27&orderby=LastUpdated&format=l

  Then, the script will be making a REST API query that looks something like this: https://rt.server.com/REST/1.0/search/ticket?query=Queue+%3D+%27RT-Queue-Name%27+AND+LastUpdated+%3E+%27-5+days%27&orderby=LastUpdated&format=l

  Where the RT Queue name in this instance is RT-Queue-Name and this query polls all RT tickets that were updated in the past 5 days.

• url_ticket_display_prefix: The URL prefix to display RT tickets, given a Ticket ID.

  For example if this value is set to something like: https://rt.server.com/Ticket/Display.html?id=

  Then, if the script processes a Ticket ID (42), it will append that number to the end of the URL when creating the corresponding JIRA ticket, like: https://rt.server.com/Ticket/Display.html?id=42

• last_fetched_timestamp: If running the script for the first time, leave this value blank. If you want to have the script re-process older RT tickets, you can clear out this value and increase the 5 day value in the search query URI listed in api_search_suffix.

[jira]

• api_url_prefix: This is the URL of JIRA web interface.

• project: When a new JIRA ticket is created by the script, it will be added to the specified project.

• component: When a new JIRA ticket is created by the script, it will be added with the specified component.

[sanity]

• reviewed: You have to change this setting from False to True once all other INI settings are defined.

TODO

• Syslog support
• Create links between related JIRA tickets
• Advance state tracking (meaning, when an RT ticket is resolved, then the correpsonding JIRA ticket should be resolved)
• Auto-add relevant watchers to the JIRA tickets
• Any attached files to RT tickets should also be attached to the JIRA ticket
• When a JIRA ticket is created, figure out some way to reply to the RT thread, indicating that a ticket has been created with the corresponding URL
• Omit requester name from summary; instead, figure out a way to update the Reporter of the JIRA ticket
• It would be nice of the ticket summary/description were built from a configurable template (eventually)
• When multiple RT tickets are submitted with the exact same subject line, treat them as a single JIRA ticket if and only if the RT Ticket IDs are +/- within 10 of each other (for intelligent clustering)

FAQ

1. rt2jira appears to be combining unrelated RT tickets into the same JIRA ticket; how can I fix this?

   Confirm that both RT tickets have an identical Subject. If that is the case, then adjust the ticket_id_correlation_range value by reducing it from 10 to some smaller value -- perhaps 2 or 1. Keep in mind, that the lower you set this value, the less likely the script will correlate/combine any RT tickets in the future.

2. I made a mistake; how do I reset the state of this script?

   Clear out the last_fetched_timestamp value so that it is completely empty. Next, adjust the 5 value in api_search_suffix to be a much higher value, in order to go further back in time. Once fixed, re-run the script and it should start processing much older RT tickets. If it still isn't going back far enough, increase the 5 value to be even higher.

3. How does this compare with other RT-to-JIRA migration scripts?

   Unlike other scripts, this script is designed to be run on a continual basis, rather than just as a one-time operation. Sure, you can use it to bulk-convert all RT tickets over to JIRA and then retire the RT instance, but this script is also flexible enough to support continual sync between RT and JIRA, so that both systems remain in operation.

4. I'm encountering the following error when running your script: HTTP 400: "The issue type selected is invalid."

   This likely means that the default issue type 'Bug' isn't valid for your project. Please edit the config.ini file and change the issue_type field from 'Bug' to some other valid issue type for your project (e.g., 'Task').