Teleport Blog - Achieving SOC2 Compliance for Teleport Cloud with Teleport On-Prem - Aug 10, 2021
Achieving SOC2 Compliance for Teleport Cloud with Teleport On-Prem
Teleport has been instrumental in helping our clients achieve difficult security and compliance requirements, and today we are proud to announce that our Cloud offering is now SOC2 Type II compliant.
Last year our on-premises product was SOC2 Type II certified, and we published an overview on our blog helping explain what SOC2 is and why it has become table stakes for B2B SaaS companies.
In this blog post, we would like to share how Teleport helped us get SOC2 compliant (we use it internally of course) and how you can achieve the same at your organization to make both developers and corporate security professionals happy.
Teleport not only helps make securing access to your servers, Kubernetes clusters, web applications, and databases easier but also makes audits go much smoother too, thanks to strong integrations with Single Sign-On (SSO) and identity providers, robust audit capabilities, access approval workflows, and change management via Terraform.
Teleport at Teleport
Internally we run a Teleport Enterprise high-availability cluster on AWS to manage access to all of our infrastructure and we integrate it with our Single Sign-On provider (Okta), our SIEM Panther , and incident management platform (PagerDuty).
At Teleport we have eliminated SSH keys (except for break glass ones) on all production infrastructure. At first, our auditors balked at this response that no users have SSH keys and we don't have to worry about revoking any because they are all short-lived, but after a quick lesson on how certificates work, they were very impressed.
SOC2 has a few control areas related to SSH management. Teleport goes beyond what's typically required, and we hope that SOC2 requirements will eventually be updated to make using certificates the expectation.
|SOC2 SSH Requirements||Teleport|
|Production access is limited to users with a valid SSH key.||Short-lived certificates are provided.|
|SSH users use unique accounts to access production Machines.||Identity-aware roles are granted with very granular access controls.|
|MFA is required for systems with customer data.||MFA can be checked during SSO/IdP authentication and on a per-session basis with Teleport Session Controls.|
|Access to production systems is revoked within a day of termination.||Certificates are short-lived, and Teleport users are federated with SSO/IdP so no additional action is needed to revoke access.|
We use Okta internally, and our process to onboard/offboard users to Teleport is simply creating/suspending users in Okta which happens automatically when our HRIS (Rippling) triggers the process. Even role changes are zero-touch because the HRIS system updates Okta's directory which then updates Teleport roles, revoking access to the systems the employee had in their prior role.
Change management and approvals
One process that often slows down developers is having to get approvals for access. It often involves creating a Jira ticket, bugging a manager to approve it, and then getting Ops to grant the access. This process is simply too slow to be effective in an incident (which is why many orgs end up bending the rules). Additionally, it's not very good security. If a single Ops engineer has the ability to grant access to anything, what happens if an Ops team member gets compromised?
With Teleport, Ops work disappears and opening tickets becomes invisible to the user thanks to approval plug-ins. The audit process is a breeze because the built-in reporting can be used with familiar tools like Jira or PagerDuty.
Making changes to role configuration is also a process that often involves opening a ticket, describing the changes to be made, getting approval, and then actually making the change. Again, it's a slow process that is a bad security practice because there is no mechanism to prevent any role change from being made which creates a vector for a hacker to escalate their permissions. With the Teleport Terraform provider this process is easy. Just open a Pull Request, get a code review and merge. We showed our auditors our merge history and GitHub approvals, and we were done.
We integrate Teleport with Okta which greatly simplifies the developer onboarding/offboarding process and is now entirely zero-touch thanks to the HRIS integration and Terraform.
Our HRIS system creates/updates and suspends users in Okta. Their user attributes are populated with HR data about their team/department that we can use in Role-Based Access Control (RBAC) rules to determine what directory groups to place them in. (Of course, we Terraformed these rules too).
The Okta group membership is then passed to Teleport during the SAML assertion informing Teleport what role to grant them. The entire process from HRIS to SSO/IdP to Teleport has built-in change control, so no single person can make an access change. HR changes require approval and Okta RBAC rule changes require a code review before Terraform applies. With this setup, we don't need to grant admin roles in Okta or Teleport since changes are made via Terraform. This provides robust security where we don't need to worry about a single user being compromised leading to privilege escalation.
Providing evidence to auditors is as simple as providing the Okta audit logs that we already share for showing that access to other systems was revoked appropriately. Now managing access and compliance for servers is as simple as SaaS app management.
Our Teleport logs get pulled into Panther SIEM using an AWS serverless process that pipes DynamoDB events into S3 to be read by Panther. Within the SIEM platform we can use SQL to query logs for audits and threat investigations and also set up alerts when suspicious activity occurs. Since we federate our users with Okta, one of our indicators of compromise is manually modifying a user.
I'm also happy to share that we now have FluentD support for log forwarding so people running all versions of Teleport can export their logs.
What key SOC2 controls does Teleport help achieve?
Teleport helps with 4 of the 9 control areas. Here is an overview:
CC6 Control Activities
Teleport helps with separation of duties using RBAC and restricts access to authorized users
- Provide role-based access controls (RBAC) using short-lived certificates and your existing identity management service.
CC6 Physical and Logical Access controls
Teleport issues temporarily security credentials according to user's role
CC7 System Operations
Teleport helps audit and monitor access.
- Audit events and session recordings are securely stored in a vault to prevent tampering.
- Convert logins, executed commands, deployments, and other events into structured audit logs.
- Monitor, share, and join interactive sessions in real time from the CLI or browser.
CC8 Change Management
Teleport helps users elevate their permissions during incidents while RBAC helps limit the need for approvals. The Teleport Slack integration allows for managers to quickly approve temporary SSH access requests.
- Let engineers request elevated permissions on the fly without ever leaving the terminal.
- Approve or deny permission requests with ChatOps workflow via Slack or other supported platforms.
- Extend and customize permission elevation workflow with a simple API and extendable plugin system.
Teleport cybersecurity blog posts and tech news
Every other week we'll send a newsletter with the latest cybersecurity news and Teleport updates.
Teleport is designed to help satisfy tough compliance requirements, and we are committed to continuing to use Teleport internally so we can continue to improve our experience for corporate security managers. Want to chat more? join us on our community slack, or start a GitHub discussion.
Stay up-to-date with the newest Teleport releases by subscribing to our monthly updates.