Running Teleport?

Teleport role mappings or trusted clusters denying access? We clear the backlog.

Teleport is excellent until a trusted cluster will not finish shaking hands, an internal.kubernetes_groups trait does not map the way the docs implied, or SSO starts rejecting valid users because a role lost its allowed logins. We read the audit log, the auth service log and the role definitions until the right people can get in and the wrong ones cannot.

  • Since 2015Fixing technology
  • FreeEstimates
  • US teamReal engineers

teleport-auth.log

still failing

What it is

Teleport at a glance

Teleport is an identity-aware access proxy for SSH servers, Kubernetes clusters, databases, Windows desktops, internal web apps and cloud APIs. It issues short-lived certificates bound to an identity, records every session for audit, and replaces long-lived SSH keys and database passwords with role-based access.

Teleport Inc. (formerly Gravitational) maintains the project with an open-source community edition and commercial Enterprise and Cloud editions. Common deployments include regulated environments that need session audit, remote teams that want to avoid VPNs, and platforms that have outgrown ad-hoc bastion hosts and shared credentials.

Type
Identity-aware access proxy for SSH, Kubernetes, databases, web apps and desktops
First released
2016
License
AGPL-3.0 community edition, proprietary Enterprise and Cloud editions
Maintained by
Teleport Inc.
Typically used for
Replacing SSH keys and VPNs, auditing privileged sessions, identity-gated access to infrastructure

The honest version

Teleport pros and cons

No technology is all upside. This is where Teleport earns its place and where it tends to cause trouble.

Where Teleport is strong

  • One identity across SSH, K8s and databasesEngineers log in once with SSO and get short-lived certs for every resource their role permits, which retires most of the shared-credential inventory a company accumulates.
  • Session recording and auditSSH, Kubernetes exec and desktop sessions are recorded with search and playback, which closes a compliance gap that is otherwise expensive to build.
  • Short-lived credentials by defaultCertificates are issued per login with a lifetime measured in hours, which neutralises most of the risk around leaked keys or stolen credentials.
  • Works for multi-cloud and on-prem togetherTrusted clusters link leaf clusters across accounts, clouds and on-premise data centres under one identity and audit trail.
  • No more VPN for most accessTeleport's proxy makes resources reachable without exposing them directly, which lets a lot of teams retire the VPN they never liked operating.
  • Honest API and Terraform supportRoles, users and clusters can be managed as code, which matters when the access policy is an auditable artefact rather than a UI click trail.

Where Teleport bites

  • Role syntax is unforgivingAllowed logins, labels, kubernetes_groups and database_users all have to line up for access to work, and a missing wildcard or a mistyped trait silently denies instead of warning.
  • Trusted clusters are hard to troubleshootHandshake failures and role-mapping surprises between root and leaf clusters produce terse errors, and the docs assume you already understand how the trust is modelled.
  • Upgrades need planningAuth, proxy and node versions have to stay within a supported window, and skipping majors or mismatching Enterprise and open-source components causes outages that are not fun to recover from.
  • Resource appetite is realThe auth service, proxies and audit storage add up at scale, particularly if you record everything and keep it for a long time.
  • AGPL and commercial mix is confusingCommunity, Enterprise and Cloud editions have different feature sets, different support levels and different licensing, and vendors sometimes end up deployed on the wrong edition for the use case.
  • SSO integration edge casesGroup claim shape, nameID format and attribute mapping vary between identity providers, and a slightly different claim produces a very polite permission-denied with no obvious remediation.

Sound familiar?

Common Teleport problems we fix

The Teleport problems that bring people to us, and what you have once each one is fixed.

access to node is denied to role(s)

Users get access denied despite having a role

A role whose allow block matches namespace and labels but has no matching login, or an SSO user whose traits never populated because the group claim had the wrong name.

Role allow rules complete and matching

trusted cluster handshake timed out

A trusted cluster will not finish trust setup

Clock skew, a self-signed root cert that the leaf cannot verify, or a proxy address that is unreachable from the leaf, with the handshake hanging until it gives up.

Trust established cleanly

the token does not allow "Kube|Node|App|DB" role

Nodes will not join the cluster

A join token that was created for a different role than the agent is trying to register as, or a token that has expired, so the agent cannot complete enrollment.

Join token scoped correctly

SSO assertion missing required group claim

SSO users cannot sign in

An identity provider that is not releasing the group claim Teleport expects, or an attribute mapping that uses a slightly different name, so every login looks like a group-less user and gets denied.

Claim mapping aligned with the IdP

Kubernetes impersonation denied

kubectl works for some users and not others

A role that grants Kubernetes access but does not populate kubernetes_groups or kubernetes_users, so the API server receives an impersonation attempt with nothing to impersonate as.

Role traits produce real Kubernetes identities

database access denied: no matching database_users

Database access is blocked at the proxy

A role that lists database_labels but forgets database_users or database_names, with Teleport refusing the connection because it cannot resolve who the engineer should connect as.

Database role fields completed

x509: certificate signed by unknown authority

TLS trust is broken between components

A proxy, agent or trusted cluster using a certificate chain the auth service does not trust, often after a rotation or a redeploy that swapped CA material.

Trust bundles synchronized

What we do

Teleport help, from rescue to clean access policy

Call us for one role that keeps denying, or hand over the whole access layer and the identity provider feeding it.

Rescue
Users locked out, trusted clusters that will not connect, or an SSO change that broke access overnight. We stabilise the proxy first and sort the policy second.
Role design
Allow and deny rules, labels, kubernetes_groups, database_users and traits arranged so least privilege is actually enforced and audits can tell what it does.
SSO integration
SAML or OIDC with the common identity providers configured so group claims flow into roles reliably, with test logins that prove it works.
Trusted clusters
Multi-cluster deployments across clouds and regions set up with correct role mapping and TLS trust, so one identity reaches all of them.
Audit and compliance
Session recording to S3 or similar, audit event forwarding to your SIEM, and retention policies that match the regulations you actually fall under.
Managed Tech
Rolling upgrades, cluster health monitoring, role change review and a human to call when an SSO change blocks a team on a Monday morning.

How it works

From first message to fixed

  1. Look Free estimate

    Tell us what's going wrong. We look at it, then tell you plainly what's wrong and what it will cost.

  2. Fix

    We fix the cause, not the symptom. If something is down, getting it back up comes first.

  3. Harden

    Backups, monitoring, updates and access, so the same problem can't quietly come back.

  4. Hand back or stay on

    You get a plain-language write-up. Or we stay on for hosting, upkeep and new work.

Often found together

Running Teleport alongside something else?

We work on the whole stack, so the problem can be wherever it actually is.

Questions

Teleport questions

Something else? Call+1 737 727 7327

Why do users get access denied in Teleport?

Almost always a role that matches the resource labels but not the login, or an SSO user whose traits never populated because of a mismatched group claim. The audit log for that login attempt names the role and the match predicates, which usually points at the missing piece quickly.

How do trusted clusters actually work?

A leaf cluster trusts a root cluster, and a root-cluster role maps to a leaf-cluster role through role-mapping rules in the trust configuration. If the mapping does not produce a role with valid logins for the leaf, access fails politely. We walk the mapping end to end when it does not.

Should we use the open-source or Enterprise edition?

Open source covers SSH, Kubernetes, databases and web apps with GitHub-style SSO. Enterprise adds SAML, OIDC, access requests, device trust and support contracts. For regulated environments or larger teams, Enterprise usually pays for itself; for small teams the community edition is fine.

Can Teleport replace our VPN?

For engineer access to servers, Kubernetes, databases and internal web apps, yes, and this is one of the most common moves. For device-level network access that your office tools still assume, you may need to keep a VPN for those specific paths until they move to identity-aware access.

How much does Teleport help cost?

The quote itself is free. Describe what the cluster is protecting and what is denying, and we return a price before work begins.

Free estimate

Teleport giving you trouble?

Tell us what you're running and what's going wrong. A person reads it and comes back with an estimate.

  • No charge to find out what it will take
  • If it's urgent, say so

Replies come from a person. By sending, you authorize us to email you back. No marketing.