GCP Memorystore Redis

Google Cloud Memorystore for Redis

Google Cloud Memorystore for Redis is a fully managed, in-memory data store service built on Redis. It provides a robust and high-performance caching layer that integrates seamlessly with Google Cloud applications.

Design Decisions

We have made several design decisions to optimize the use of Google Cloud Memorystore for Redis:

• High Availability: Optional support for high availability using STANDARD_HA tier, which includes read replicas for enhanced resilience.
• Private Service Access: Configure connections to Memorystore through private service access for enhanced security.
• Version Management: Supports multiple Redis versions with seamless upgrades.
• Monitoring and Alarms: Includes built-in alarms for critical metrics like memory usage.
• Authentication and Security: Redis AUTH is enabled for securing access to the instance.

Runbook

Checking Instance Status

If you're experiencing connectivity or performance issues, the first step is to check the status of your Redis instance.

gcloud redis instances describe [INSTANCE_ID] --region=[REGION]

You should see details about the instance's status, memory usage, and network connectivity. Look for the state field to confirm that the instance is operational.

Viewing Log Entries

To help diagnose issues, you can check the log entries associated with your Redis instance.

gcloud logging read "resource.type=redis_instance AND resource.labels.instance_id=[INSTANCE_ID]" --limit 30

This will return the latest log entries related to your Redis instance, which may contain error messages or other diagnostic information.

Redis Performance Issues

If your Redis instance is experiencing slow performance, you can use the redis-cli tool to connect to the instance and run diagnostic commands.

redis-cli -h [HOST] -p [PORT] -a [AUTH_STRING]

Once connected, you can run the following commands to check performance:

• Checking Memory Usage:

  info memory
  

  Look for the used_memory and used_memory_peak fields to understand how much memory is being consumed.

• Checking Slow Queries:

  slowlog get 10
  

  This retrieves the last 10 queries that took the longest time to execute. If you see many slow queries, consider optimizing your queries or upgrading your instance size.

High Memory Usage Alerts

If you've received an alert for high memory usage, it may be necessary to clear some cache or upgrade your instance.

• Flushing All Data: Keep in mind this will remove all data from the Redis instance.

  redis-cli -h [HOST] -p [PORT] -a [AUTH_STRING] FLUSHALL
  

• Evicting Keys Based on Policy: Configure an eviction policy if one isn't already set to manage memory usage gracefully.

  redis-cli -h [HOST] -p [PORT] -a [AUTH_STRING] config set maxmemory-policy allkeys-lru
  

Connectivity Issues

If you experience connectivity issues:

• Check Authorized Networks: Ensure your VPC network is authorized to connect to the Redis instance.

  gcloud redis instances describe [INSTANCE_ID] --region=[REGION] --format="get(authorizedNetwork)"
  

• Ping the Redis Endpoint: Verify you can reach the Redis endpoint from your application.

  ping [HOST]