IBM Cloud inactive identities: Ideas for automated processing

Common cleanup is a part of all account administration and safety finest practices, not only for cloud environments. In our blog post on identifying inactive identities, we seemed on the APIs provided by IBM Cloud Id and Entry Administration (IAM) and tips on how to make the most of them to acquire particulars on IAM identities and API keys. Some readers offered suggestions and requested on tips on how to proceed and act on recognized inactive identities.

In response, we’re going lay out attainable steps to take. We present tips on how to discover and revoke current privileges and what to think about. Furthermore, we talk about how the totally different id sorts will be faraway from an account. We additionally present some instructions on tips on how to script and probably automate these administrative duties:

Recap: Inactive identities

IBM Cloud Id and Entry Administration (IAM) helps different forms of identities. They embody customers and repair IDs—each with related API keys—in addition to trusted profiles. When such an id or an related API key has not been used to authenticate for a set time, it’s thought-about inactive.

IBM Cloud IAM supplies functionality to create reports on inactive identities. By default, identities are thought-about inactive after they haven’t logged in or been in use in 30 days. When making a report by using the API or an SDK, you may specify different time frames (e.g., 90 days).

Inactive identities pose a safety threat as a result of they may be not maintained and be simpler to assault. To enhance safety, you must revoke entry privileges from inactive identities and possibly even solely take away them from the cloud account.

There’s, nevertheless, an operational threat with particular identities which might be solely used for quarterly or annual processing (which, in our opinion, is dangerous safety design). If cleaned up, their related duties could fail. This state of affairs may very well be addressed by preserving tabs on how inactive identities and their privileges are cleaned up.

Automated cleanup

Performing on found inactive identities may very well be accomplished manually, however needs to be automated for effectivity and improved safety. Each guide and automatic cleanup might comply with a course of like this:

1. Generate and retrieve a report on inactive identities for the specified date vary.
2. Test the reported identities towards an inventory of exempted IDs.
3. Loop over every non-exempted id and remove it from all IBM Cloud IAM access groups. Additionally, be sure that no directly granted permissions exist.
4. Go over discovered API keys and delete them.

For all steps, log the findings and actions taken for audit and enhancements.

Relying in your company insurance policies, you may need to clear up month-to-month or quarterly. When triggering the report generation in step one, you may specify the period (the vary in hours) for what to think about as inactive. To keep away from the danger of shutting down vital identities, you must preserve an inventory or database with identities which might be excluded from cleanup (Step 2 above). That listing is also used to tell apart between totally different insurance policies like month-to-month or quarterly checks.

When processing every discovered inactive id (e.g., customers, service IDs, trusted profiles), it’s pretty straightforward to revoke assigned privileges. IBM Cloud IAM supplies a REST API with a DELETE to remove an IAM identity from all associated access groups (Step 3 above, see screenshot beneath).

If following finest practices, permissions ought to solely be assigned by way of entry teams and never instantly. You may confirm this rule by retrieving the list of directly granted privileges for the IAM identity. If such a privilege (entry administration coverage) is discovered, there’s an API to delete that policy (Step 3). You may see our weblog put up “IBM Cloud security: How to clean up unused access policies” for extra info.

The report on inactive identities additionally features a part on API keys. API keys are related to both a consumer or service ID. The query is how quickly to scrub them up by deleting the API key. Much like eradicating privileges from an id, deleting an related API key could break purposes. Determine what’s finest to your cloud atmosphere and meets company requirements.

The above cleanup steps will be scripted and run manually. You can additionally automate the cleanup by taking an method much like what we describe on this blog post on automated data scraping. Use IBM Cloud Code Engine with a cron subscription to set off execution on set dates or intervals:

Customers, service IDs and trusted profiles

Above, we mentioned tips on how to revoke privileges from inactive identities. To additional clear up the account and improve safety, you must contemplate deleting unused service IDs and trusted profiles and eradicating customers from the account. These actions may very well be a follow-up after stripping permissions—when it’s clear that these identities not are wanted. Moreover, you may periodically list all users and verify their states. Take away customers out of your account which have an invalid, suspended or (sort of) deleted state.

IBM Cloud has API features to remove a user from an account, to delete a service ID and its associated API keys and to delete a trusted profile.

Conclusions

Common account cleanup is a part of account administration and safety finest practices, not only for cloud environments. In our blog post on identifying inactive identities, we seemed on the APIs provided by IBM Cloud Id and Entry Administration (IAM) and tips on how to make the most of them to acquire particulars on IAM identities and API keys.

On this weblog put up, we mentioned an method on tips on how to routinely clear up privileges that have been granted to now inactive identities. It is very important word that some housekeeping within the type of (audit) logs and an inventory of exempted identities is required to maintain your apps and workloads working. In that sense, do it, however don’t overdo it.

See these weblog posts and repair documentation for additional info:

In case you have suggestions, options, or questions on this put up, please attain out to me on Twitter (@data_henrik), Mastodon (@data_henrik@mastodon.social) or LinkedIn.