atlassian

Atlassian Demo Data Seeder

This module provisions Atlassian structure (Jira projects, Ops teams/schedules) and invokes a deterministic seeder to generate narrative-shaped demo data for Developer Health analytics.

What this creates

• 10 Jira Software projects
• 10 Ops teams with weekly on-call rotations
• ~12k Jira issues across 24 months
• Incidents + postmortem follow-ups (labeled and linked)
• Sprint history and spillover (when Agile APIs are available)
• out/manifest.json with distribution summaries

Requirements

• Jira Cloud + Jira Software
• Atlassian Ops (JSM Ops/On-call) access for the same site
• API token with project admin access
• Python 3 with pip3 (PyYAML and requests will be auto-installed by Terraform)

Environment variables

Set the following via terraform.tfvars or environment variables:

• jira_url (e.g. https://your-domain.atlassian.net)
• jira_user (email)
• jira_token
• atlassian_cloud_id
• atlassian_domain
• atlassian_org_id
• project_lead_account_id (optional; defaults to current admin user)
• enable_project_creation (optional; set to false if projects already exist)
• team_member_account_ids (list; optional)
• generated_user_count (optional)
• generated_user_domain (optional)
• enable_user_creation (optional)

Optional toggles:

• enable_issue_creation (default false - set to true to create issues)
• enable_sprints (default true)
• enable_transitions (default true)
• disable_incidents (default false - set to true to skip incidents)
• enable_comments (default false)

Usage

cd atlassian
terraform init
terraform apply

To run in dry-run mode (manifest only):

# terraform.tfvars
# enable_issue_creation defaults to false, so dry-run is the default behavior
# To actually create issues, set:
enable_issue_creation = true

Reset / destroy

terraform destroy

The seeder is idempotent for issues: it uses a deterministic external id stored as a label (extid-<hash>). Re-running terraform apply skips previously seeded issues.

Rate limits and retries

The seeder retries failed API calls up to 3 times with a short sleep. If you hit rate limits, re-run terraform apply after a few minutes.

Permissions needed

• Project admin for Jira project creation
• Issue create + transition for Jira and JSM projects
• Ops team and schedule management access

Notes on timestamps

Jira Cloud does not allow backdating created/resolved without import permissions. This seeder stores simulated timestamps in the issue property seed_meta and in out/manifest.json for analytics.

Files

• seed/story_map.yaml controls project/team mapping and arc distributions
• seed/seed_jira.py contains the deterministic data generator
• out/manifest.json is produced after seeding