How do I turn on log delivery in an ElastiCache for Redis OSS or ElastiCache for Valkey cluster?

Resolution

Prerequisites:

• Create an AWS Identity and Access Management (IAM) policy that has the required permissions to publish to Amazon CloudWatch Logs or Amazon Data Firehose.
• Attach the policy to the appropriate IAM user or role.

The IAM policy must have the following permissions:

• logs:CreateLogDelivery
• logs:UpdateLogDelivery
• logs:DeleteLogDelivery
• logs:GetLogDelivery
• logs:ListLogDeliveries

Note: Valkey 7.x and later and Redis OSS cache clusters and replication groups that use engine version 6.0 and later support slow log. Valkey 7.x, and Redis OSS cache clusters and replication groups that use engine version 6.2 and later support engine log.

Turn on log delivery for your cluster

You can use ElastiCache APIs to turn on and configure log delivery when you create or modify a cluster. ElastiCache delivers log entries to the specified destination in either JSON or text format.

To turn on log delivery for Redis when you create or modify a cluster, see Specifying log delivery using the Console.

To turn on log delivery for Redis slow log, you must periodically retrieve a fixed number of slow log entries from the engine. The entries must be between the following two parameters that you define in the parameter group:

• slowlog-log-slower-than
• slowlog-max-len

When a command exceeds the slowlog-log-slower-than threshold, a new entry is added to the slow log. The slowlog-max-len configuration determines the maximum number of entries for the slow log.

NOTE: If no commands that take longer than slowlog-log-slower-than, then no slow logs are generated.

When you turn on log delivery for Redis engine log, the logs record internal operations of the Redis engine and provide additional insights. Insights include initiation of backup and critical events, such as failed synchronization.

When you turn on engine logs during cluster creation, you can view engine log entries that relate to successful synchronization between the primary and replica nodes.

However, when you turn on engine logs after cluster creation, the log stream might be empty because there's no specific activity at the Redis engine.

To confirm that you correctly configured Amazon CloudWatch Logs for your cluster, create a manual backup of your cluster. If the logs of the backup are sent to your CloudWatch log group, then you correctly configured CloudWatch Logs.

Related information

Log type and log format specifications

Contents of a slow log entry

Contents of an engine log entry