Skip to main content

Cloud-Managed AWS RDS and Aurora MySQL Automatic Installation

Use this instruction to install Releem Agent automatically via CloudFormation to AWS Fargate, or manually on an EC2 instance. Releem Agent will run in a container on AWS Fargate or directly on the EC2 instance, depending on your choice.

Releem supports Amazon RDS MySQL, Amazon RDS MariaDB, and Amazon Aurora MySQL. Configure one Agent for each DB instance endpoint.

Parameter groups​

To apply recommended configuration automatically, sign in to your AWS account and prepare custom parameter groups. AWS-managed default groups cannot be modified and are not valid onboarding targets.

Amazon RDS MySQL or MariaDB (non-Aurora):

  • Create a custom DB parameter group for your database version, for example releem-agent.
  • Attach it to the instance as DB parameter group.
  • Leave the cluster parameter group empty.
  • Reboot the instance if AWS requires it.

Amazon Aurora MySQL:

  • Create a custom DB parameter group for the instance and a custom DB cluster parameter group for the cluster.
  • Attach both groups in AWS. The names you configure for Releem must match the groups attached in AWS.
  • Both groups must be ready before the Agent applies recommendations. A recommendation can target either group.
  • Only the Agent that targets the writer modifies cluster parameters. That Agent needs rds:ModifyDBClusterParameterGroup permission.

The requirements for the RDS or Aurora instance:

  • Enhanced Monitoring to collect system performance metrics.

  • Performance Insights to collect MySQL performance metrics.

  • Performance Schema to collect MySQL performance metrics. Change the following variables in the custom parameter group and reboot the database instance if required:

    performance_schema=ON
    slow_query_log=ON

You have the following options to install Releem Agent:

The CloudFormation will deploy Releem Agent container using AWS Fargate service to collect metrics from your AWS RDS instance and send them to Releem Cloud Platform. The template will create roles to run Releem Agent with the following permissions:

  • logs:Get*
  • rds:Describe*
  • cloudwatch:Get*
  • ecr:GetAuthorizationToken
  • ecr:BatchCheckLayerAvailability
  • ecr:GetDownloadUrlForLayer
  • ecr:BatchGetImage
  • rds:ModifyDBParameterGroup
  • rds:ModifyDBClusterParameterGroup

To move forward just do the following steps:

  1. Create read-only user "releem" in MySQL using the instruction. To enable Automatic SQL Query Optimization please add Additional Permissions.

  2. Open the CloudFormation Quick Create page in your browser, check that the correct region is selected, and fill in the following fields:

    • Image: Releem Agent Docker image in the releem/releem-agent:[version] format. Use the latest version available on Docker Hub.
    • APIKey: Releem API Key from the Profile page in the Releem Customer Portal.
    • DBID: RDS DB Instance ID. For Aurora, set the instance name you want to monitor, not the cluster endpoint. Deploy one stack per instance.
    • DatabaseType: leave mysql (default) for RDS MySQL, MariaDB, and Aurora MySQL.
    • DBUser: User to collect metrics from the SQL server, that we added in the step 1.
    • DBPassword: Password for user
    • DBSSLMode: set true when the instance requires SSL.
    • SecurityGroupIDs: Releem Agent SecurityGroups to connect to RDS. Please Allow All Outbound traffic for Releem Agent Security Group.
    • SubnetIDs: Releem Agent container should be able to connect to RDS.
    • QueryOptimization: set True if Releem Agent should collect additional information for Automatic SQL Query Optimization.
    • DBParameterGroup: Custom DB parameter group attached to the instance. Required for RDS and Aurora instance parameters.
    • DBClusterParameterGroup: Custom DB cluster parameter group attached to Aurora. Required for Aurora. Leave empty only for non-Aurora RDS.

DatabaseType defaults to mysql and exposes MySQL (DB_*) credentials to the container. Select postgresql only for Aurora PostgreSQL; see PostgreSQL.

  1. Next, click "Create Stack".

Common Issues for AWS RDS​

To check Releem Agent logs please open CloudWatch -> Log Groups and then select Releem Agent log group.

  1. Failed to read log stream %s:%s: %s RDSOSMetrics

    • Enable Enhanced monitoring for your RDS instance.

  2. No Latency graph on Releem Dashboard

    • Enable Performance Insights and Performance Schema for your RDS instance.

  3. Error 1045 (28000): Access denied for user 'releem'@'' (using password: YES)

    • Check password for Releem User that you created according to this guide. And reinstall Releem Agent.
    • To set new password please run in MySQL console the following command: 
      ALTER USER 'releem'@'%' IDENTIFIED BY 'New-Password-Here';

  4. Connect: connection timed out

    • Check that SecurityGroup of RDS instance has Inbound rule to accept connections from Releem Agent installed in ECS.

  5. CloudFormation stack is not finished, still "in progress" for a long time

    • Enable Performance Insights and Enhanced monitoring for your RDS instance.
    • Check that SecurityGroup that you set for Releem Agent has a rule for All outbound traffic.

  6. Error 1142 (42000): SELECT command denied to user 'releem'@'[serverIP]' for table 'events_statements_history'

  7. Performance_schema_ variables aren't being applied on AWS RDS

    • Amazon RDS only applies performance_schema_* variables when performance_schema is explicitly set to 1 in the Parameter Group. By default, it's set to 0, even though Performance Insights enables it automatically at the database level.
    • Solution: Set performance_schema from 0 to 1 in your Parameter Group, save the changes, and reboot your RDS instance.

  8. Aurora apply fails because the cluster parameter group is missing or does not match

    • Create a custom DB cluster parameter group, attach it to the cluster, and set DBClusterParameterGroup / AWS_RDS_CLUSTER_PARAMETER_GROUP to that exact name.
    • AWS-managed default groups cannot be modified.

  9. Cluster parameters are not applied on a reader

    • Only the Agent targeting the writer modifies cluster parameters. Keep an Agent on the writer instance.