Skip to content settings

DETAILS: Tier: Free, Premium, Ultimate Offering:

This page contains information about the settings that are used on, available to GitLab SaaS customers.

See some of these settings on the instance configuration page of

Email confirmation has the:

Password requirements has the following requirements for passwords on new accounts and password changes:

  • Minimum character length 8 characters.
  • Maximum character length 128 characters.
  • All characters are accepted. For example, ~, !, @, #, $, %, ^, &, *, (), [], _, +, =, and -.

SSH key restrictions uses the default SSH key restrictions.

SSH host keys fingerprints

Go to the current instance configuration to see the SSH host key fingerprints on

  1. Sign in to GitLab.
  2. On the left sidebar, select Help ({question-o}) > Help.
  3. On the Help page, select Check the current instance configuration.

In the instance configuration, you see the SSH host key fingerprints:

Algorithm MD5 (deprecated) SHA256
ECDSA f1:d0:fb:46:73:7a:70:92:5a:ab:5d:ef:43:e2:1c:35 SHA256:HbW3g8zUjNSksFbqTiUWPWg2Bq1x8xdGUrliXFzSnUw
ED25519 2e:65:6a:c8:cf:bf:b2:8b:9a:bd:6d:9f:11:5c:12:16 SHA256:eUXGGm1YGsMAS7vkcx6JOJdOGHPem5gQp4taiCfCLB8
RSA b6:03:0e:39:97:9e:d0:e7:24:ce:a3:77:3e:01:42:09 SHA256:ROQFvPThGrW4RuWLoL9tq9I9zJ42fK4XywyRtbOz/EQ

The first time you connect to a repository, one of these keys is displayed in the output.

SSH known_hosts entries

Add the following to .ssh/known_hosts to skip manual fingerprint confirmation in SSH: ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIAfuCHKVTjquxvt6CM6tdG4SLp1Btn/nOeHHE5UOzRdf ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAABAQCsj2bNKTBSpIYDEGk9KxsGh3mySTRgMtXL583qmBpzeQ+jqCMRgBqB98u3z++J1sKlXHWfM9dyhSevkMwSbhoR8XIq/U0tCNyokEi/ueaBMCvbcTHhO7FcwzY92WK4Yt0aGROY5qX2UKSeOvuP4D6TPqKF1onrSzH9bx9XUf2lEdWT/ia1NEKjunUqu1xOB/StKDHMoX4/OKyIzuS0q/T1zOATthvasJFoPrAjkohTyaDUz2LN5JoH839hViyEG82yB+MjcFV5MU3N1l1QL3cVUCh93xSaua1N85qivl+siMkPGbO5xR/En4iEY6K2XPASUEMaieWVNTRCtJ4S8H+9 ecdsa-sha2-nistp256 AAAAE2VjZHNhLXNoYTItbmlzdHAyNTYAAAAIbmlzdHAyNTYAAABBBFSMqzJeV9rUzU4kWitGjeR4PWSa29SPqJ1fVkhtj3Hw9xjLVXVYrU9QlYWrOLXBpQ6KWjbjTDTdDkoohFzgbEY=

Mail configuration sends emails from the domain by using Mailgun, and has its own dedicated IP addresses:


The IP addresses for are subject to change at any time.

Service Desk alias email address

On, there's a mailbox configured for Service Desk with the email address: contact-project+%{key} To use this mailbox, configure the custom suffix in project settings.


See our backup strategy.

To back up an entire project on, you can export it either:

With exports, be aware of what is and is not included in a project export.

GitLab is built on Git, so you can back up just the repository of a project by cloning it to another computer. Similarly, you can clone a project's wiki to back it up. All files uploaded after August 22, 2020 are included when cloning.

Delayed group deletion

DETAILS: Tier: Premium, Ultimate Offering:

After May 08, 2023, all groups have delayed deletion enabled by default.

Groups are permanently deleted after a seven-day delay.

If you are on the Free tier, your groups are immediately deleted, and you will not be able to restore them.

You can view and restore groups marked for deletion.

Delayed project deletion

DETAILS: Tier: Premium, Ultimate Offering:

After May 08, 2023, all groups have delayed project deletion enabled by default.

Projects are permanently deleted after a seven-day delay.

If you are on the Free tier, your projects are immediately deleted, and you will not be able to restore them.

You can view and restore projects marked for deletion.

Inactive project deletion

Inactive project deletion is disabled on

Alternative SSH port can be reached by using a different SSH port for git+ssh.

Setting Value
Port 443

An example ~/.ssh/config is the following:

  User git
  Port 443
  PreferredAuthentications publickey
  IdentityFile ~/.ssh/gitlab

GitLab Pages

Some settings for GitLab Pages differ from the defaults for self-managed instances:

Domain name
IP address
Support for custom domains {check-circle} Yes
Support for TLS certificates {check-circle} Yes
Maximum site size 1 GB
Number of custom domains per GitLab Pages website 150

The maximum size of your Pages site depends on the maximum artifact size, which is part of GitLab CI/CD.

Rate limits also exist for GitLab Pages.

GitLab container registry

Domain name
IP address
CDN domain name
CDN IP address

To use the GitLab container registry, Docker clients must have access to:

  • The registry endpoint and for authorization.
  • Google Cloud Storage or Google Cloud Content Delivery Network to download images. is fronted by Cloudflare. For incoming connections to, you must allow CIDR blocks of Cloudflare (IPv4 and IPv6).

GitLab CI/CD

Below are the current settings regarding GitLab CI/CD. Any settings or feature limits not listed here are using the defaults listed in the related documentation.

Setting Default (self-managed)
Artifacts maximum size (compressed) 1 GB See Maximum artifacts size.
Artifacts expiry time From June 22, 2020, deleted after 30 days unless otherwise specified (artifacts created before that date have no expiry). See Default artifacts expiration.
Scheduled Pipeline Cron */5 * * * * See Pipeline schedules advanced configuration.
Maximum jobs in active pipelines 500 for Free tier, 1000 for all trial tiers, 20000 for Premium, and 100000 for Ultimate. See Number of jobs in active pipelines.
Maximum CI/CD subscriptions to a project 2 See Number of CI/CD subscriptions to a project.
Maximum number of pipeline triggers in a project 25000 for all tiers See Limit the number of pipeline triggers.
Maximum pipeline schedules in projects 10 for Free tier, 50 for all paid tiers See Number of pipeline schedules.
Maximum pipelines per schedule 24 for Free tier, 288 for all paid tiers See Limit the number of pipelines created by a pipeline schedule per day.
Maximum number of schedule rules defined for each security policy project Unlimited for all paid tiers See Number of schedule rules defined for each security policy project.
Scheduled job archiving 3 months (from June 22, 2020). Jobs created before that date were archived after September 22, 2020. Never.
Maximum test cases per unit test report 500000 Unlimited.
Maximum registered runners Free tier: 50 per-group / 50 per-project
All paid tiers: 1000 per-group / 1000 per-project
See Number of registered runners per scope.
Limit of dotenv variables Free tier: 50 / Premium tier: 100 / Ultimate tier: 150 See Limit dotenv variables.
Authorization token duration (minutes) 15 To set a custom value, in the Rails console, run ApplicationSetting.last.update(container_registry_token_expire_delay: <integer>), where <integer> is the desired number of minutes.
Maximum downstream pipeline trigger rate (for a given project, user, and commit) 350 per minute See Maximum downstream pipeline trigger rate.

Package registry limits

The maximum file size for a package uploaded to the GitLab package registry varies by format:

Package type
Conan 5 GB
Generic 5 GB
Helm 5 MB
Maven 5 GB
npm 5 GB
NuGet 5 GB
Terraform 1 GB

Account and limit settings has the following account limits enabled. If a setting is not listed, the default value is the same as for self-managed instances:

Setting default
Repository size including LFS 10 GB
Maximum import size 5 GiB
Maximum export size 40 GiB
Maximum remote file size for imports from external object storages 10 GiB
Maximum download file size when importing from source GitLab instances by direct transfer 5 GiB
Maximum attachment size 100 MiB
Maximum decompressed file size for imported archives 25 GiB
Maximum push size 5 GiB

If you are near or over the repository size limit, you can either:

NOTE: git push and GitLab project imports are limited to 5 GiB per request through Cloudflare. Imports other than a file upload are not affected by this limit. Repository limits apply to both public and private projects.

Default import sources

The import sources that are available to you by default depend on which GitLab you use:

  • all available import sources are enabled by default.
  • GitLab self-managed: no import sources are enabled by default and must be enabled.

IP range uses the IP ranges and for traffic from its Web/API fleet. This whole range is solely allocated to GitLab. You can expect connections from webhooks or repository mirroring to come from those IPs and allow them. is fronted by Cloudflare. For incoming connections to, you might need to allow CIDR blocks of Cloudflare (IPv4 and IPv6).

For outgoing connections from CI/CD runners, we are not providing static IP addresses. All instance runners are deployed into Google Cloud Platform (GCP) in us-east1. Any IP-based firewall can be configured by looking up IP address ranges or CIDR blocks for GCP.

Hostname list

Add these hostnames when you configure allow-lists in local HTTP(S) proxies, or other web-blocking software that governs end-user computers. Pages on load content from these hostnames:

  • *
  • *
  • *
  • *

Documentation and Company pages served over and also load certain page content directly from common public CDN hostnames.


The following limits apply for webhooks.

Rate limits

The number of times a webhook can be called per minute, per top-level namespace. The limit varies depending on your plan and the number of seats in your subscription.

Plan Default for
Free 500
Premium 99 seats or fewer: 1,600
100-399 seats: 2,800
400 seats or more: 4,000
Ultimate and open source 999 seats or fewer: 6,000
1,000-4,999 seats: 9,000
5,000 seats or more: 13,000

Other limits

Setting Default for
Number of webhooks 100 per project, 50 per group (subgroup webhooks are not counted towards parent group limits )
Maximum payload size 25 MB
Timeout 10 seconds
Multiple Pages deployments 100 extra deployments (Premium tier), 500 extra deployments (Ultimate tier)

For self-managed instance limits, see:

Runner SaaS

Runner SaaS is the hosted, secure, and managed build environment you can use to run CI/CD jobs for your hosted project.

For more information, see Runner SaaS.

Puma uses the default of 60 seconds for Puma request timeouts.

Maximum number of reviewers and assignees

Merge requests enforce these maximums:

  • Maximum assignees: 200
  • Maximum reviewers: 200 rate limits

NOTE: See Rate limits for administrator documentation.

When a request is rate limited, GitLab responds with a 429 status code. The client should wait before attempting the request again. There are also informational headers with this response detailed in rate limiting responses.

The following table describes the rate limits for

Rate limit Setting
Protected paths for an IP address 10 requests per minute
Raw endpoint traffic for a project, commit, or file path 300 requests per minute
Unauthenticated traffic from an IP address 500 requests per minute
Authenticated API traffic for a user 2,000 requests per minute
Authenticated non-API HTTP traffic for a user 1,000 requests per minute
All traffic from an IP address 2,000 requests per minute
Issue creation 200 requests per minute
Note creation on issues and merge requests 60 requests per minute
Advanced, project, or group search API for an IP address 10 requests per minute
GitLab Pages requests for an IP address 1,000 requests per 50 seconds
GitLab Pages requests for a GitLab Pages domain 5,000 requests per 10 seconds
GitLab Pages TLS connections for an IP address 1,000 requests per 50 seconds
GitLab Pages TLS connections for a GitLab Pages domain 400 requests per 10 seconds
Pipeline creation requests for a project, user, or commit 25 requests per minute
Alert integration endpoint requests for a project 3,600 requests per hour
Pull mirroring intervals 5 minutes
API requests from a user to /api/v4/users/:id 300 requests per 10 minutes
GitLab package cloud requests for an IP address (introduced in GitLab 16.11) 3,000 requests per minute

More details are available on the rate limits for protected paths and raw endpoints.

GitLab can rate-limit requests at several layers. The rate limits listed here are configured in the application. These limits are the most restrictive per IP address. For more information about the rate limits for, see an overview.

Rate limiting responses

For information on rate limiting responses, see:

Protected paths throttle responds with HTTP status code 429 to POST requests at protected paths that exceed 10 requests per minute per IP address.

See the source below for which paths are protected. This includes user creation, user confirmation, user sign in, and password reset.

User and IP rate limits includes a list of the headers responded to blocked requests.

See Protected Paths for more details.

IP blocks

IP blocks can occur when receives unusual traffic from a single IP address that the system views as potentially malicious. This can be based on rate limit settings. After the unusual traffic ceases, the IP address is automatically released depending on the type of block, as described in a following section.

If you receive a 403 Forbidden error for all requests to, check for any automated processes that may be triggering a block. For assistance, contact GitLab Support with details, such as the affected IP address.

Git and container registry failed authentication ban responds with HTTP status code 403 for 1 hour, if 30 failed authentication requests were received in a 3-minute period from a single IP address.

This applies only to Git requests and container registry (/jwt/auth) requests (combined).

This limit:

  • Is reset by requests that authenticate successfully. For example, 29 failed authentication requests followed by 1 successful request, followed by 29 more failed authentication requests would not trigger a ban.
  • Does not apply to JWT requests authenticated by gitlab-ci-token.

No response headers are provided.

git requests over https always send an unauthenticated request first, which for private repositories results in a 401 error. git then attempts an authenticated request with a username, password, or access token (if available). These requests might lead to a temporary IP block if too many requests are sent simultaneously. To resolve this issue, use SSH keys to communicate with GitLab.

Pagination response headers

For performance reasons, if a query returns more than 10,000 records, GitLab excludes some headers.

Visibility settings

Projects, groups, and snippets have the Internal visibility setting disabled on

SSH maximum number of connections defines the maximum number of concurrent, unauthenticated SSH connections by using the MaxStartups setting. If more than the maximum number of allowed connections occur concurrently, they are dropped and users get an ssh_exchange_identification error.

Group and project import by uploading export files

To help avoid abuse, the following are rate limited:

  • Project and group imports.
  • Group and project exports that use files.
  • Export downloads.

For more information, see:

Non-configurable limits

See non-configurable limits for information on rate limits that are not configurable, and therefore also used on Gitaly RPC concurrency limits

Per-repository Gitaly RPC concurrency and queuing limits are configured for different types of Git operations such as git clone. When these limits are exceeded, a fatal: remote error: GitLab is currently unable to handle this request due to load message is returned to the client.

For administrator documentation, see limit RPC concurrency. logging

We use Fluentd to parse our logs. Fluentd sends our logs to Stackdriver Logging and Cloud Pub/Sub. Stackdriver is used for storing logs long-term in Google Cold Storage (GCS). Cloud Pub/Sub is used to forward logs to an Elastic cluster using pubsubbeat.

You can view more information in our runbooks such as:

Job logs

By default, GitLab does not expire job logs. Job logs are retained indefinitely, and can't be configured on to expire. You can erase job logs manually with the Jobs API or by deleting a pipeline. at scale

In addition to the GitLab Enterprise Edition Linux package install, uses the following applications and settings to achieve scale. All settings are publicly available, as Kubernetes configuration or Chef cookbooks.

Elastic cluster

We use Elasticsearch and Kibana for part of our monitoring solution:


We use Fluentd to unify our GitLab logs:


Prometheus complete our monitoring stack:


For the visualization of monitoring data:


Open source error tracking:


Service discovery:


High Performance TCP/HTTP Load Balancer:

Sidekiq runs Sidekiq as an external process for Ruby job scheduling.

The current settings are in the Kubernetes pod configuration.