summaryrefslogtreecommitdiff
path: root/doc/user/group/saml_sso/scim_setup.md
blob: 64ac4f15d5f0db6e39d96005ee577fc163ae7f04 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
---
type: howto, reference
---

# SCIM provisioning using SAML SSO for Groups **(SILVER ONLY)**

> [Introduced](https://gitlab.com/gitlab-org/gitlab-ee/merge_requests/9388) in [GitLab.com Silver](https://about.gitlab.com/pricing/) 11.10.

System for Cross-domain Identity Management (SCIM), is an open standard that enables the
automation of user provisioning. When SCIM is provisioned for a GitLab group, membership of
that group is synchronized between GitLab and the identity provider.

GitLab's [SCIM API](../../../api/scim.md) implements part of [the RFC7644 protocol](https://tools.ietf.org/html/rfc7644).

Currently, the following actions are available:

- CREATE
- UPDATE
- DELETE (deprovisioning)

The following identity providers are supported:

- Azure

## Requirements

- [Group SSO](index.md) needs to be configured.
- The `scim_group` feature flag must be enabled:

  Run the following commands in a Rails console:

  ```sh
  # Omnibus GitLab
  gitlab-rails console

  # Installation from source
  cd /home/git/gitlab
  sudo -u git -H bin/rails console RAILS_ENV=production
  ```

  To enable SCIM for a group named `group_name`:

  ```ruby
  group = Group.find_by_full_path('group_name')
  Feature.enable(:group_scim, group)
  ```

## GitLab configuration

Once [Single sign-on](index.md) has been configured, we can:

1. Navigate to the group and click **Settings > SAML SSO**.
1. Click on the **Generate a SCIM token** button.
1. Save the token and URL so they can be used in the next step.

![SCIM token configuration](img/scim_token.png)

## Identity Provider configuration

### Azure

First, double check the [Single sign-on](index.md) configuration for your group and ensure that **Name identifier value** (NameID) points to `user.objectid` or another unique identifier. This will match the `extern_uid` used on GitLab.

![Name identifier value mapping](img/scim_name_identifier_mapping.png)

#### Set up admin credentials

Next, configure your GitLab application in Azure by following the
[Provisioning users and groups to applications that support SCIM](https://docs.microsoft.com/en-us/azure/active-directory/manage-apps/use-scim-to-provision-users-and-groups#provisioning-users-and-groups-to-applications-that-support-scim) 
section in Azure's SCIM setup documentation.

During this configuration, note the following:

- The `Tenant URL` and `secret token` are the ones retrieved in the
  [previous step](#gitlab-configuration).
- Should there be any problems with the availability of GitLab or similar
  errors, the notification email set will get those.
- It is recommended to set a notification email and check the **Send an email notification when a failure occurs** checkbox.
- For mappings, we will only leave `Synchronize Azure Active Directory Users to AppName` enabled.

You can then test the connection by clicking on **Test Connection**. If the connection is successful, be sure to save your configuration before moving on. See below for [troubleshooting](#troubleshooting).

#### Configure attribute mapping

1. Click on `Synchronize Azure Active Directory Users to AppName`, to configure the attribute mapping.
1. Click **Delete** next to the `mail` mapping.
1. Map `userPrincipalName` to `emails[type eq "work"].value` and change it's **Matching precedence** to `2`.
1. Map `mailNickname` to `userName`.
1. Create a new mapping by clicking **Add New Mapping** then set **Source attribute** to `objectId`, **Target attribute** to `id`, **Match objects using this attribute** to `Yes`, and **Matching precedence** to `1`.
1. Create a new mapping by clicking **Add New Mapping** then set **Source attribute** to `objectId`, and **Target attribute** to `externalId`.
1. Click the `userPrincipalName` mapping and change **Match objects using this attribute** to `No`.

    Save your changes and you should have the following configuration:

   ![Azure's attribute mapping configuration](img/scim_attribute_mapping.png)

    NOTE: **Note:** If you used a unique identifier **other than** `objectId`, be sure to map it instead to both `id` and `externalId`.

1. Below the mapping list click on **Show advanced options > Edit attribute list for AppName**.
1. Leave the `id` as the primary and only required field.

   NOTE: **Note:**
   `username` should neither be primary nor required as we don't support
   that field on GitLab SCIM yet.

   ![Azure's attribute advanced configuration](img/scim_advanced.png)

1. Save all the screens and, in the **Provisioning** step, set
   the `Provisioning Status` to `On`.

    ![Provisioning status toggle switch](img/scim_provisioning_status.png)

   NOTE: **Note:**
   You can control what is actually synced by selecting the `Scope`. For example,
   `Sync only assigned users and groups` will only sync the users assigned to
   the application (`Users and groups`), otherwise, it will sync the whole Active Directory.

Once enabled, the synchronization details and any errors will appear on the
bottom of the **Provisioning** screen, together with a link to the audit logs.

## Troubleshooting

### Testing Azure connection: invalid credentials

When testing the connection, you may encounter an error: **You appear to have entered invalid credentials. Please confirm you are using the correct information for an administrative account**. If `Tenant URL` and `secret token` are correct, check whether your group path contains characters that may be considered invalid JSON primitives (such as `.`). Removing such characters from the group path typically resolves the error.

### Azure: (Field) can't be blank sync error

When checking the Audit Logs for the Provisioning, you can sometimes see the
error `Namespace can't be blank, Name can't be blank, and User can't be blank.`

This is likely caused because not all required fields (such as first name and 
last name) are present for all users being mapped.

As a workaround, try an alternate mapping:

1. Follow the Azure mapping instructions from above.
1. Delete the `name.formatted` target attribute entry.
1. Change the `displayName` source attribute to have `name.formatted` target attribute.