Files
coder/docs/tutorials/persistent-shared-workspaces.md
T
Nick Vigilante 81c40579dc feat: enable Coder.StepLength Vale rule (#30791)
Adds `Coder.StepLength`, a Vale `script` rule for [Keep steps
short](https://github.com/coder/coder/blob/main/docs/.style/style-guide/procedural-writing.md#keep-steps-short).
It flags an ordered-list step whose instruction, the step's first
paragraph, has more than 20 words, counted as [How to count
words](https://github.com/coder/coder/blob/main/docs/.style/style-guide/voice-and-tone.md#how-to-count-words)
describes.
Paragraphs below the instruction explain the step and aren't counted.

- Severity: `warning`.
- Scope: `docs/**` except `docs/.style/style-guide/**`.
- Baseline: 124 findings in 60 hand-written files.
The cleanup shortens each long instruction, moves its explanation into
one paragraph below it, or splits a step that holds several actions.
After cleanup `vale docs/` reports 0 `Coder.StepLength` findings, and
the other rules report no new findings on the changed files.
- The guide section now says the limit covers the instruction, and that
one action or one explanation must not be broken across paragraphs to
meet it.
- Generated pages (`docs/reference/{api,cli,agent-api}`,
`configuration-reference.md`, `prometheus.md`): no findings, so no
`.vale.ini` exclusion is needed.

Linear: DOCS-1272

<details><summary>Decision log</summary>

**Review change (Ian's review)**

- The first version limited every paragraph of a step to 20 words.
The cleanup for it chopped explanations into several short paragraphs,
which made steps harder to read and sometimes changed their meaning.
- The rule now counts only the first paragraph segment of each step.
The content files were reset to `main` and the cleanup was redone
against the narrower rule.
- Cleanup preferences, in order: shorten the instruction; keep a short
instruction and put the explanation, with its original wording, in one
paragraph below it; split a step into one step for each action; for
descriptive lists (such as "How it works"), keep each item one
statement.
- `docs/admin/users/oidc-auth/microsoft.md` keeps the split into
separate steps, with Ian's wording suggestions applied.
"click" became "select" in step 16 because `Coder.SelectClick` flags it.

**Rule shape**

- A `script` rule (Tengo) with `scope: raw`, because the count depends
on Markdown structure that per-paragraph scopes hide.
- A step is an ordered-list item with a `1.` or `1)` marker.
  A nested ordered item is its own step.
  Bullets, even under a step, are not steps.
- The instruction ends at a blank line, a fence, a nested list, a table
row, a blockquote, block-level HTML, or an explicit hard break (`<br>`,
a trailing backslash, or two trailing spaces).
- Fenced code blocks, HTML comments, tables, images, and nested bullets
add no words.
- CRLF: `\r` counts as whitespace, and offsets come from the raw text.

**Counting approximations (also in the rule header)**

- A code span and a `**bold**` or `__bold__` span are one word each.
- Words split on whitespace, so a hyphenated compound and
`30&nbsp;seconds` are one word.
- Link text counts word by word, destinations are skipped, images add no
words, and `<https://...>` autolinks are one word.
- A token with no letter or digit, such as the `>` in `**A** > **B**`,
is not a word.
- Tengo indexes strings by rune but measures them in bytes, so the
script rewrites each non-ASCII rune to ASCII of the same byte length
first.

**Tests** (fixtures under `/tmp/sltest/`, not committed; run with the
repo `.vale.ini`, LF and CRLF)

- Flagged: a long instruction, a short action and a long result in one
paragraph, a long nested ordered step.
- Not flagged: a long paragraph below a short instruction, text after a
`<br>` hard break, a long paragraph after a code block, a long nested
bullet.
- LF and CRLF output is identical.
- Pre-commit hooks (`fmt/markdown`, `lint/markdown`, `lint/emdash`,
`lint/style-claims`, `lint/typos`) pass.
- Not run on Windows or macOS.
The rule is plain text processing, and the CRLF cases simulate Windows
line endings on Linux.

</details>

*Generated by [Coder Agents](https://coder.com/docs/ai-coder/agents) on
behalf of @nickvigilante.*
2026-10-26 20:43:38 -04:00

8.8 KiB

title
title
Persistent shared workspaces with service accounts

Note

This guide requires a Premium license because service accounts are a Premium feature. For more details, contact your account team.

This guide walks through setting up a long-lived workspace that is owned by a service account and shared with a rotating set of users. Because no single person owns the workspace, it persists across team changes and every user authenticates as themselves.

This pattern is useful for any scenario where a workspace outlives the people who use it:

  • On-call rotations: Engineers share a workspace pre-loaded with runbooks, dashboards, and monitoring tools. Access rotates with the shift schedule.
  • Shared staging or QA: A team workspace hosts a persistent staging environment. Testers and reviewers are added and removed as sprints change.
  • Pair programming: A service-account-owned workspace gives two or more developers a shared environment without either one owning (and accidentally deleting) it.
  • Contractor onboarding: An external team gets scoped access to a workspace for the duration of an engagement, then access is revoked.

The steps below use an on-call SRE workspace as a running example, but the same commands apply to any of the scenarios above. Substitute the usernames, group names, and template to match your use case.

Prerequisites

  • A running Coder deployment (v2.32+) with workspace sharing enabled. Sharing is on by default for OSS; Premium deployments may require admin configuration.
  • The Coder CLI installed and authenticated.
  • An account with the Owner or User Admin role.
  • OIDC authentication configured so shared users log in with their corporate SSO identity. Configure refresh tokens to prevent session timeouts during long work sessions.
  • A wildcard access URL configured (for example, *.coder.example.com) so that shared users can access workspace apps without a 404.
  • (Recommended) IdP Group Sync configured if your identity provider manages group membership for the teams that will share the workspace.

1. Create a service account

Create a dedicated service account that will own the shared workspace. Service accounts are non-human accounts intended for automation and shared ownership. Because no individual user owns the workspace, there are no personal credentials to expose and the shared environment is not affected when any user leaves the team or the organization.

# On-call example — substitute a name that fits your use case
coder users create \
  --username oncall-sre \
  --service-account

2. Generate an API token for the service account

Generate a long-lived API token so you can create and manage workspaces on behalf of the service account:

coder tokens create \
  --user oncall-sre \
  --name oncall-automation \
  --lifetime 8760h

Store this token securely (for example, in a secrets manager like Vault or AWS Secrets Manager).

Important

Never distribute this token to end users. The token is for workspace administration only. Shared users authenticate as themselves and reach the workspace through sharing.

3. Create the workspace

Authenticate as the service account and create the workspace:

export CODER_SESSION_TOKEN="<token-from-step-2>"

coder create oncall-sre/oncall-workspace \
  --template your-oncall-template \
  --use-parameter-defaults \
  --yes

Tip

Design a dedicated template for the workspace with the tools your team needs pre-installed (for example, monitoring dashboards for on-call, test runners for QA). Set subdomain = true on workspace apps so that shared users can access web-based tools without a 404. See Access workspace apps in shared workspaces.

4. Share the workspace

Use coder sharing share to grant access to users who need the workspace:

coder sharing share oncall-sre/oncall-workspace --user alice

This gives alice the default use role, which allows connection via SSH and workspace apps, starting and stopping the workspace, and viewing logs and stats.

To grant admin permissions (which includes all use permissions as well as renaming, updating, and inviting others to join with the use role):

coder sharing share oncall-sre/oncall-workspace --user alice:admin

To share with multiple users at once:

coder sharing share oncall-sre/oncall-workspace --user alice:admin,bob

To share with an entire Coder group:

coder sharing share oncall-sre/oncall-workspace --group sre-oncall

Note

Groups can be synced from your identity provider using IdP Sync. If your IdP already manages team membership, sharing with a group is the simplest approach.

5. Rotate access

When team membership changes, remove outgoing users and add incoming ones:

# Remove outgoing user
coder sharing remove oncall-sre/oncall-workspace --user alice

# Add incoming user
coder sharing share oncall-sre/oncall-workspace --user carol

Important

The workspace must be restarted for user removal to take effect.

Verify current sharing status at any time:

coder sharing status oncall-sre/oncall-workspace

For every subcommand and flag, see the coder sharing reference.

6. Automate access changes (optional)

For use cases with frequent rotation (such as on-call shifts), you can integrate the share/remove commands into external tooling like PagerDuty, Opsgenie, or a cron job.

Rotation script

#!/bin/bash
# rotate-access.sh
# Usage: ./rotate-access.sh <outgoing-user> <incoming-user>

WORKSPACE="oncall-sre/oncall-workspace"
OUTGOING="$1"
INCOMING="$2"

if [ -n "$OUTGOING" ]; then
  echo "Removing access for $OUTGOING..."
  coder sharing remove "$WORKSPACE" --user "$OUTGOING"
fi

echo "Granting access to $INCOMING..."
coder sharing share "$WORKSPACE" --user "$INCOMING"

echo "Restarting workspace to apply changes..."
coder restart "$WORKSPACE" --yes

echo "Current sharing status:"
coder sharing status "$WORKSPACE"

Group-based rotation with IdP Sync

If your identity provider manages group membership (for example, an sre-oncall group in Okta or Azure AD), you can skip manual share/remove commands entirely:

  1. Configure Group Sync to synchronize the group from your IdP to Coder.

  2. Share the workspace with the group once:

    coder sharing share oncall-sre/oncall-workspace --group sre-oncall
    
  3. When your IdP rotates group membership, Coder group membership updates on next login.

    All current members have access; removed members lose access after a workspace restart.

Find shared workspaces

Shared users can find workspaces shared with them:

# List all shared workspaces you can access, including your own
coder list --search shared:true

# List workspaces shared with you
coder list --search shared_with_user:me

# List workspaces shared with a specific user
coder list --search shared_with_user:alice

# List workspaces shared with a specific group
coder list --search shared_with_group:sre-oncall

Troubleshooting

Shared user sees 404 on workspace apps

Workspace apps using path-based routing block non-owners by default. Configure a wildcard access URL and set subdomain = true on the workspace app in your template.

Removed user still has access

Access removal requires a workspace restart. Run coder restart <workspace> after removing a user or group.

Group sync not updating membership

Group membership changes in your IdP are not reflected until the user logs out and back in. Group sync runs at login time, not on a polling schedule. Check the logs from coderd, the process that runs the control plane, with CODER_LOG_FILTER=".*userauth.*|.*groups returned.*" for details. See Troubleshooting group sync for more information.

Next steps