Rackspace says attackers used leaked credentials to access its Spot cloud platform and act as legitimate users in some customer organizations. They could create Kubernetes pods, read cluster Secrets, and list nodes and storage volumes. Secrets can hold passwords, API keys and tokens for services outside the cluster, extending the recovery work beyond Spot itself.

Evidence status: vendor-reported intrusion, documented in a customer notification obtained by BleepingComputer. Lawrence Abrams’s November 21 report quotes the notice. The reviewed evidence does not establish how many organizations were affected or which downstream credentials were subsequently abused.

Cluster access exposed other services’ credentials

Rackspace’s notice places the unauthorized access on November 17, approximately 00:20–22:00 UTC. The company says a customer report alerted it on November 18, when it confirmed that customer’s organization was affected. The notice describes leaked credentials as the entry point without explaining how they originally escaped.

Kubernetes Secrets are stored credential objects, not a guarantee that credentials remain safe after an authorized identity is compromised. Rackspace warns that an attacker could have read any Secret in the affected user’s clusters. Spot API tokens and refresh tokens held by that user were also potentially exposed.

That scope is conditional on the affected identity and organization. The report does not show that every Spot cluster was accessed, or that listing a storage volume meant reading all its contents.

Rackspace says it revoked affected tokens and sessions, removed unauthorized pods, rotated keys, and strengthened access controls and monitoring. Its administration console was unavailable during containment, temporarily obstructing customers’ ability to manage clusters and rotate credentials.

A November 21 service-restoration post by Sirish Raghuram says Spot was fully restored. Abrams identifies Raghuram as a Platform9 co-founder involved in Spot’s development and operation. That is an operator’s recovery statement, not independent proof that credentials copied earlier are harmless.

Replace credentials where they are accepted

For organizations Rackspace identifies as affected, its notification recommends:

  • Reissue Spot API tokens and replace automation refresh tokens. Update the jobs that use them and retire the old values.
  • Download fresh Kubernetes configuration files. Replace affected access material in administrator and automation workflows through established account channels.
  • Rotate credentials stored in cluster Secrets. Replace or revoke them at the database, registry, cloud or other service that accepts them, then update dependent workloads. Editing a cluster object alone cannot invalidate a copied credential elsewhere.
  • Review downstream service logs from November 17 onward. Rackspace specifically names databases, container registries, Git repositories, cloud storage and third-party services. Preserve relevant records while investigating possible use of exposed credentials.

Rackspace also supplied network and suspicious-object indicators in the notification. Use those alongside identity and workload history; a different address or pod name cannot establish that an affected environment was untouched.

The operational boundary is the credential’s reach. Restoring Spot access enables recovery, while revocation and log review address what an intruder could carry away.

Caveat: the incident details rely on Rackspace’s customer notice; BleepingComputer had received no answer on affected-customer totals or confirmed data theft when it published.