Managing Stale Repos and Branches

[Chas Berndt]

Is there a way to determine what branches and repositories are not being used so we can slate them for decommission? When we migrated from GitLab to Stash we had nearly 100 repos that were little labs that people created and just left untouched. 

I'd like to have a way to find repositories or branches that have not had commits to them in the past, say six months. Is there a tool for this or has anyone used Stash's APIs to accomplish something like this?

[Marcin]

This can be done with the REST API and a script - basically the steps are:

• retrieve all the repositories
• for each repository get the branches ordered by modification date (most recent first)
• get the latest commit on the top branch - that should be the most recent commit to the repo

I've put together a small Python script that does this - the only dependency is the Requests library for easier HTTP. You'll need to change some variables at the top to match your environment.

import requests
import datetime
 
STASH_URL = 'http://localhost:7990/stash'
USERNAME = 'admin'
PASSWORD = 'admin'
auth = (USERNAME, PASSWORD)

THRESHOLD = datetime.datetime.now() - datetime.timedelta(weeks=26) # only show repos with commits older than half a year ago

repos = requests.get(STASH_URL + '/rest/api/1.0/repos?limit=9999', auth=auth)
values = repos.json()['values']
print("Searching %d repos..." % len(values))
for repo in values:
  repo = (repo['project']['key'], repo['slug'])
  branches_request = requests.get(STASH_URL + '/rest/api/1.0/projects/%s/repos/%s/branches?orderBy=MODIFICATION&limit=1' % repo, auth=auth)
  branches = branches_request.json()['values']
  if len(branches) > 0:
    most_recent_branch = branches[0]
    most_recent_commit = most_recent_branch['latestChangeset']
    commit_request = requests.get(STASH_URL + '/rest/api/1.0/projects/%s/repos/%s/commits/%s' % (repo + (most_recent_commit,)), auth=auth)
    commit = commit_request.json()
    commit_date = datetime.datetime.fromtimestamp(commit['authorTimestamp'] / 1000)
    if commit_date <= THRESHOLD:
      print("Most recent commit in repo %s in project %s was at %s by %s (%s)" % (repo + (commit_date, commit['author']['name'], commit['message'])))

[Akshata Nayak]

The branches_request list does not support 'values' key . It is throwing the error it key should be an integer not a string. Stuck there.

[suryatej yaramada]

Will this script works for Bitbucket to find stale branches and unused Bitbucket repos ? we are using Bitbucket cloud