MongoDB configuration

Configuring MongoDB storage is basically just creating an user, which has permissions to read and write to a database.

When using MongoDB atlas you setup your project. By default you should have (or you should be asked to add) a user, which is in role of a admin.

mongodb-users

It is recommended to add another user, who can write to any database: mongodb-new

After the user is created you need to create a file secret.yaml inside cfg directory. Make sure only owner can read this file as it will contain username, password and server to the MongoDB database.

secret.yaml structure

The file can contain configuration for multiple project, the main section is same as the name of your project (e.g. hello-world). To setup connection to a db server create section databaseThis section can contain several options but all of them are passed to the construcotr of the python’s pymongo.mongo_client.MongoClient constructor. Please refer to api.mongodb.com for further information.

Take a look at the example of the secret.yaml file here.

If your MongoDB server is not hosted, you must setup MongoDB authorization (for example via /etc/mongodb.conf):

# /etc/mongodb.conf
net:
  bindIp: 0.0.0.0
  port:   27017

security:
  authorization: enabled

secret.yaml examples

1. Example 1 (single host):

   hello-world:
     database:
       host:           [mongodb.server.example.com:27017]
       connect:        true
       authSource:     admin
       username:       writer
       password:       password-here
   

2. Example 2 (mongodb configuration with 3 hosts):

   hello-world:
     database:
       host:
         - cluster0-shard-00-00-foobar.mongodb.net:27017
         - cluster0-shard-00-01-foobar.mongodb.net:27017
         - cluster0-shard-00-02-foobar.mongodb.net:27017
       replicaSet:     Cluster0-shard-0
       connect:        true
       authSource:     admin
       authMechanism:  SCRAM-SHA-1
       ssl:            true
       username:       writer
       password:       password-here
   

If secret.yaml is setup properly, you can easily collect benchmark data to a database. By default the data will be saved into a database with the name of your project name. You can change this behaviour by adding another section to the secret.yaml:

hello-world:
  artifacts:
    db_name:          customDbName      # name of the database to save date to
    
    col_timers_name:  customRepColName  # name of the collection which will contain
                                        # benchmark data (reports, frames)
    
    col_files_name:   customFSColName   # name of the collection which will contain
                                        # files and logs when error during run occured

Note: Assuming we are trying to override artifacts location fot the project hello-world