Platform.sh allows you to completely define and configure the topology and services you want to use on your project.
Unlike other PaaS services, Platform.sh is batteries included which means that you don't need to subscribe to an external service to get a cache or a search engine. And that those services are managed. When you back up your project, all of the services are backed-up.
Services are configured through the
.platform/services.yaml file you will need to commit to your Git repository. This section describes specifics you might want to know about for each service.
If you don't have a
.platform folder, you need to create one:
mkdir .platform touch .platform/services.yaml
Here is an example of a
database1: type: mysql:10.1 disk: 2048 database2: type: postgresql:9.6 disk: 1024
name you want to give to your service. You are free to name each service as you wish (lowercase alphanumeric only).
WARNING: Because we support multiple services of the same type (you can have 3 different MySQL instances), changing the name of the service in
services.yaml will be interpreted as destroying the existing service and creating a new one. This will make all the data in that service disappear forever. Remember to always back up your environment in which you have important data before modifying this file.
type of your service. It's using the format
If you specify a version number which is not available, you'll see this error when pushing your changes:
Validating configuration files. E: Error parsing configuration files: - services.mysql.type: 'mysql:5.6' is not a valid service type.
Service types and their supported versions include:
||1.2, 1.3, 1.7|
||2.1, 2.2, 2.3, 2.4|
||10.0, 10.1, 10.2, 10.3, 10.4|
||1.4, 1.5, 1.6|
||3.0, 3.2, 3.4, 3.6|
||9.6, 10, 11, 12|
||3.5, 3.6, 3.7, 3.8|
||3.2, 4.0, 5.0|
||3.6, 4.1, 6.3, 6.6, 7.6, 7.7, 8.0, 8.4|
disk attribute is the size of the persistent disk (in MB) allocated to the service.
For example, the current default storage amount per project is 5GB (meaning 5120MB) which you can distribute between your application (as defined in
.platform.app.yaml) and each of its services. For memory-resident-only services such as
disk key is not available and will generate an error if present.
notes Currently we do not support downsizing the persistent disk of a service.
By default, Platform.sh will allocate CPU and memory resources to each container automatically. Some services are optimized for high CPU load, some for high memory load. By default, Platform.sh will try to allocate the largest "fair" size possible to all services, given the available resources on the plan. That is not always optimal, however, and you can customize that behavior on any service or on any application container. See the application sizing page for more details.
All services have their system timezone set to UTC by default. In most cases that is the best option. For some applications it's possible to change the application timezone, which will affect only the running application itself.
- MySQL - You can change the per-connection timezone by running SQL
SET time_zone = <timezone>;.
- PostgreSQL - You can change the timezone of current session by running SQL
SET TIME ZONE <timezone>;.
In order for a service to be available to an application in your project (Platform.sh supports not only multiple backends but also multiple applications in each project) you will need to refer to it in the .platform.app.yaml file which configures the relationships between applications and services.
All services offer one or more
endpoints. An endpoint is simply a named set of credentials that can be used to access the service from other applications or services in your project. Only some services support multiple user-defined endpoints. If you do not specify one then one will be created with a standard defined name, generally the name of the service type (e.g.,
solr). An application container, defined by a
.platform.app.yaml file, always exposes and endpoint named
http to allow the router to forward requests to it.
When defining relationships in a configuration file you will always address a service as
<endpoint>. See the appropriate service page for details on how to configure multiple endpoints for each service that supports it.
Once a service is running and exposed as a relationship, its appropriate credentials (host name, username if appropriate, etc.) will be exposed through the
PLATFORM_RELATIONSHIPS environment variable. The structure of each is documented on the appropriate service's page, along with sample code for how to connect to it from your application. Note that different applications manage configuration differently so the exact code will vary from one application to another.
Be aware that the keys in the
PLATFORM_RELATIONSHIPS structure are fixed but the values they hold may change on any deployment or restart. Never hard-code connection credentials for a service into your application. You should re-check the environment variable every time your script or application starts.
Access to the database or other services is only available from within the cluster. For security reasons they cannot be accessed directly. However, they can be accessed over an SSH tunnel. There are two ways to do so. (The example here uses MariaDB but the process is largely identical for any service.)
In either case, you will also need the service credentials. For that, run
platform relationships. That will give output similar to the following:
redis: - service: rediscache ip: 246.0.82.19 cluster: jyu7waly36ncj-master-7rqtwti host: redis.internal rel: redis scheme: redis port: 6379 database: - username: user scheme: mysql service: mysqldb ip: 246.0.80.37 cluster: jyu7waly36ncj-master-7rqtwti host: database.internal rel: mysql path: main query: is_master: true password: '' port: 3306
That indicates that the
database relationship can be accessed at host
user, and an empty password. The
path key contains the database name,
main. The other values can be ignored.
note When using the default endpoint on MySQL/MariaDB, the password is usually empty. It will be filled in if you define any custom endpoints. As there is only the one user and port access is tightly restricted anyway the lack of a password does not create a security risk.
The first option is to open an SSH tunnel for all of your services. You can do so with the Platform.sh CLI, like so:
$ platform tunnel:open SSH tunnel opened on port 30000 to relationship: redis SSH tunnel opened on port 30001 to relationship: database Logs are written to: ~/.platformsh/tunnels.log List tunnels with: platform tunnels View tunnel details with: platform tunnel:info Close tunnels with: platform tunnel:close
tunnel:open command will connect all relationships defined in the
.platform.app.yaml file to local ports, starting at 30000. You can then connect to those ports on
localhost using the program of your choice.
In this example, we would connect to
localhost:30001, database name
main, with username
user and an empty password.
platform tunnels command will list all open tunnels:
+-------+---------------+-------------+-----------+--------------+ | Port | Project | Environment | App | Relationship | +-------+---------------+-------------+-----------+--------------+ | 30000 | a43m75zns6k4c | master | [default] | redis | | 30001 | a43m75zns6k4c | master | [default] | database | +-------+---------------+-------------+-----------+--------------+
Alternatively, many database applications (such as MySQL Workbench and similar tools) support establishing their own SSH tunnel. Consult the documentation for your application for how to enter SSH credentials, including telling it where your SSH private key is. (Platform.sh does not support password-based SSH authentication.)
To get the values to use, the easiest way is to run
platform ssh --pipe. That will return a command line that can be used to connect over SSH, from which you can pull the appropriate information. For example:
In this case, the username is
jyu7waly36ncj-master-7rqtwti--app and the host is
ssh.us.platform.sh. Note that the host will vary per region, and the username will vary per-environment.
In this example, we would configure our database application to setup a tunnel to
ssh.us.platform.sh as user
jyu7waly36ncj-master-7rqtwti--app, and then connect to the database on host
user, empty password, and database name