How would you centralize contractor access for a small mixed server fleet?

November 20, 2026 5 answers
Asked by _r67~dovrani.4

We have around 10 Linux VMs, a few Windows servers, and contractors who rotate through projects. Offboarding means checking SSH keys and disabling accounts on multiple machines, and we also need to know who logged in where and when.

Is a central gateway worth running at this size, or is there a simpler way to manage access and logging?

Answers

9d~fablegrommet_246

Start with configuration management if the main pain is touching every Linux box. Keep keys and access rules per named user in a repository, then use Ansible to apply changes across the fleet. Removing someone means updating the source of truth and running the change, rather than hunting through each server.

That does not automatically solve Windows offboarding or give you a central login history. You still need a way to disable those accounts and collect logs. The payoff is a much simpler starting point if you are not ready to operate a gateway.

Reply to this answer
Trelvani+9c

SSH certificates are worth a look. Servers trust an internal certificate authority, and contractors use keys signed for a short period. Once someone leaves, stop issuing certificates to them. The remaining validity period matters, though. An already issued certificate can still work until it expires unless you have a separate revocation mechanism.

Vault can issue short-lived SSH credentials tied to roles. This addresses Linux access, not the whole Windows and audit problem, so treat it as one part of the setup.

2h!copperbadge_793:

Short lifetimes also mean you need a reliable way for authorized people to get new certificates. That is worth testing before relying on it for day-to-day access.

Reply to this answer
4s!gobletratchet_671

A gateway starts looking reasonable when contractors come and go often and you need a clear record of access. JumpServer, Teleport, and Guacamole are options to evaluate for a mixed SSH and RDP environment. Check the current edition and configuration against your exact requirements, especially identity integration, audit logs, and session recording.

There is a tradeoff. You are adding a service that needs updates, backups, monitoring, and a recovery plan. Also make sure disabling access at the gateway actually blocks the paths contractors use. Direct server access and separate local accounts can otherwise remain available.

_r67~dovrani.4:

Good point about direct access. A gateway would not help offboarding much if the old routes and accounts were still usable.

Reply to this answer
q.592!velquaro

A jump host is a simpler architectural step. Give contractors remote access to that host, then restrict which internal machines they can reach from it. Cutting off their jump host access removes that route to the rest of the fleet.

It still needs individual identities and useful logs, and it only works as an offboarding control if contractors cannot reach the servers through another route.

Reply to this answer
8v!casknotch_517

There is also the hosted route if maintaining another cluster sounds miserable for one admin. A service such as Tailscale may help centralize remote access, but check its current pricing and which SSH, Windows, and audit features actually fit your setup. Do not assume a network access service replaces account management on every server.

Reply to this answer

Add your answer

Answers are reviewed before publication.

Related Discussions