Add services includes many services, so you don’t have to subscribe to external cache or search engine services. Because the services are included in your project, you can manage them through Git and they’re backed up together with the rest of your project.

Your project defines the services configuration in a file named services.yaml located in the .platform directory. If you don’t need any services (such as for a static website), you don’t need to include this file in your repository.

Read on to see how to add services.


Add a service 

Adding a service is a two-step process.

1. Configure the service 

All service configuration happens in the services configuration file (.platform/services.yaml) in your Git repository.

Configure your service in the following pattern:

    # Other options...

An example service configuration for two databases might look like this:

    type: mariadb:10.5
    disk: 2048
    type: postgresql:13
    disk: 1024

This YAML file is a dictionary defining all of the services you want to use. The top-level key is a custom service name, which you use to identify the service in step 2. You can give it any name you want with lowercase alphanumeric characters, hyphens, and underscores.

Service options 

The following table presents the keys you can define for each service:

Name Type Required Description
type string Yes One of the available services in the format type:version.
disk integer For some services The size in MB of the persistent disk allocated to the service. Can’t be set for memory-resident-only services such as memcache and redis. Limited by your plan settings.
size string How many CPU and memory resources to allocate to the service. Possible values are AUTO (default), S, M, L, XL, 2XL, and 4XL. Limited by your plan settings.
configuration dictionary For some services Some services have additional specific configuration options that can be defined here, such as specific endpoints. See the given service page for more details.
relationships dictionary For some services Some services require a relationship to your app. The content of the dictionary has the same type as the relationships dictionary for app configuration. The endpoint_name for apps is always http.

You can decrease the size of an existing disk for a service. If you do so, be aware that:

  • Backups from before the downsize are incompatible and can no longer be used. You need to create new backups.
  • The downsize fails if there’s more data on the disk than the desired size.
Supported regions 

Downsizing a service’s persistent disk is available on all regions except the legacy regions:


These regions will be upgraded in the future. If you’re on one of these regions and want to downsize a disk now, migrate your project to a newer region.


Resources are distributed across all containers in a project from the total available from your plan size.

By default, allocates CPU and memory resources to each container automatically. Some services are optimized for high CPU load, some for high memory load. If your plan is sufficiently large for bigger containers, you can increase the size of your service container.

Note that service containers in development environments are always set to size S.

2. Connect the service 

Once you have configured a service, you need to create a relationship to connect it to an app. This is done in your app configuration for relationships.

The relationship follows this pattern:

An example relationship to connect to the databases given in the example in step 1:
    mysql_database: "database1:mysql"
    postgresql_database: "database2:postgresql"

As with the service name, you can give the relationship any name you want with lowercase alphanumeric characters, hyphens, and underscores. It helps if the service name and relationship name are different, but it isn’t required.

Each service offers one or more endpoints for connections, depending on the service. An endpoint is a named set of credentials to give access to other apps and services in your project. If you don’t specify one in the service configuration, a default endpoint is created. The default endpoint varies by service, generally being its type (such as mysql or solr).

Available services 

The following table presents the available service types and their versions. Add them to the type key of the service configuration in the format type:version.

Service type Supported versions
Headless Chrome chrome-headless 95, 91, 86, 84, 83, 81, 80, 73
Elasticsearch elasticsearch
InfluxDB influxdb 2.3
Kafka kafka 3.2
MariaDB/MySQL mariadb 10.6, 10.5, 10.4, 10.3
Memcached memcached 1.6, 1.5, 1.4
MongoDB mongodb
MongoDB Premium mongodb-enterprise 5.0, 4.4, 4.2
MariaDB/MySQL mysql 10.6, 10.5, 10.4, 10.3
Network Storage network-storage 2.0
OpenSearch opensearch 2, 1.2, 1.1
Oracle MySQL oracle-mysql 8.0, 5.7
PostgreSQL postgresql 14, 13, 12, 11, 10
RabbitMQ rabbitmq 3.11, 3.10, 3.9
Redis redis 7.0, 6.2
Persistent Redis redis-persistent 7.0, 6.2
Solr solr 8.11
Varnish varnish 6.3, 6.0, 5.2, 5.1
Vault KMS vault-kms 1.12

Service versions 

These services generally follow semantic versioning conventions. You choose which major/minor version to add to your project. Patch versions are applied periodically for bug fixes and the like. When you deploy your app, you always get the latest available patches.

Service timezones 

All services have their system timezone set to UTC by default. For some services, you can change the timezone for the running service (this doesn’t affect the container itself and so logs are still in UTC).

Connect to a service 

For security reasons, you can’t access services directly through HTTP. You can connect through your app or by opening an SSH tunnel to access the service directly.

When connecting to a service from an app, you may want to use one of the configuration readers. These tools make it easier to get credentials inside your app.

Alternatively, once a service is running and exposed as a relationship, its credentials (such as the host, username, and password) are available through the PLATFORM_RELATIONSHIPS environment variable. The available information is documented on each service’s page along with sample code for how to connect to it from your app.

The keys in the PLATFORM_RELATIONSHIPS variable are fixed, but the values may change on deployment or restart. So use the environment variable rather than hard coding the values.

Connecting to a service using an SSH tunnel is a two-step process.

1. Obtain service credentials 

To get the credentials for a given service, run the following command:

$ platform relationships

You get output like the following:

        username: user
        scheme: mysql
        service: database
        fragment: null
        public: false
        cluster: abcdefgh1234567-main-abcd123
        host: database.internal
        rel: mysql
            is_master: true
        path: main
        password: ''
        type: 'mariadb:10.6'
        port: 3306
        host_mapped: false
        url: 'mysql://user:@database.internal:3306/main'

With this example, you can connect to the database relationship with the user user, an empty password, and the database name main (from the path). The url property shows a full database connection that can be used from your app.

2. Open an SSH tunnel 

Open a single SSH tunnel by running the following CLI command:

platform tunnel:single --relationship RELATIONSHIP_NAME

By default, this opens a tunnel at You can specify the port for the connection using the --port flag.

You can then connect to this service in a separate terminal or locally running app. With the example above, you connect to a URL like the following: mysql://user:@'