Manage clusters
This document explains how to connect to and view a MemStore cluster and manage its monitoring, security group, parameter group, and other operational settings.
To operate a MemStore cluster reliably, regularly monitor memory usage. When memory reaches the configured maxmemory, eviction is performed according to maxmemory-policy (default: allkeys-lru).
- Valkey
maxmemory: About 66%, or two-thirds, of the instance type memory - Redis OSS
maxmemory: About 66% of available system memory
Connect to a MemStore cluster
After creating and configuring a MemStore cluster, you can connect to it with an engine client.
Prerequisites
Prepare the following environment:
- A VM (Virtual Machine) instance that uses the same network as the MemStore cluster
- A security group that permits access to the cluster endpoint
Allow the port configured when the cluster was created: the custom port in the 2000–50000 range or the default TCP 6379 port.
Connect to a cluster
-
See Create and connect to a Virtual Machine instance, connect to the instance, and prepare a Redis or Valkey client.
-
Go to the KakaoCloud Console > Data Store > MemStore.
-
In the Cluster menu, select the cluster.
-
On the Details tab, copy the Endpoint and Port.
-
Run the connection command from the VM instance.
Connect to a MemStore clusterredis-cli -h {ENDPOINT} -p {PORT}Item Description {ENDPOINT}Endpoint on the Details tab {PORT}Connection port
- Default:6379
- Or the custom port specified during cluster creation
For a cluster that uses TLS or user authentication, add the client options required by the cluster settings. You can download the CA file for TLS from the Details tab.
View the cluster list
- Go to the KakaoCloud Console > Data Store > MemStore.
- View the cluster list in the Cluster menu.
| Item | Description |
|---|---|
| Name | Name entered during creation; click it to view details |
| Status | Current cluster status; see Cluster |
| Engine version | Engine version running on the nodes |
| Node type | Node type selected during creation; hover to view vCPU and memory information |
| Cluster mode | Whether cluster mode is enabled |
| Automatic failover | Whether automatic failover is enabled for node failures - Redis OSS only |
| Automatic backup retention | Retention period; changes apply to subsequently created automatic backups |
| Shard count | Number of shards; a hyphen (-) is displayed when cluster mode is disabled- Redis OSS only |
| Node count | Number of nodes - Redis OSS only |
| Created at | Date and time when the cluster was created |
View cluster details
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, select a cluster.
- View the information on each tab.
| Tab | Description |
|---|---|
| Details | Operational status, endpoints, and other key cluster information |
| Shard/Node or Node list | Shards, nodes, and their status; see Manage shards and nodes |
| Backup | Backup information and operations; see Manage backups |
| Network/Security | Security groups attached to the cluster |
| Access control | User authentication and transport encryption settings |
| Monitoring | Monitoring information for cluster nodes |
- Details
- Shard/Node
- Backup
- Network/Security
- Access control
- Monitoring
| Item | Description |
|---|---|
| Name | Cluster name |
| ID | Cluster ID |
| Description | Cluster description |
| Operation status | Current status; see Cluster |
| Created by | User who created the cluster |
| Created at | Cluster creation date and time |
| Cluster mode | Enabled or disabled |
| Node type | Cluster node type |
| Shard | Number of shards; displayed when cluster mode is enabled |
| Node | Number of nodes and automatic failover status |
| Engine version | Cluster engine version |
| License | Engine license version |
| Parameter group | Applied group and status; see Parameter application status |
| Endpoint | Read/write endpoint when cluster mode is enabled |
| Primary endpoint | Read/write endpoint when cluster mode is disabled |
| Read endpoint | Read-only endpoint when cluster mode is disabled |
| Port | Custom port or default 6379 |
On the Backup tab, you can view backups and create, restore, copy, export, or delete a manual backup.
| Item | Description |
|---|---|
| Backup count | Total number of cluster backups |
| Automatic backup start time | Configured start time |
| Automatic backup retention | Configured retention period; changes apply to new backups |
| Name | Backup name |
| Status | Backup status; see Backup |
| Type | Automatic or manual |
| Memory at backup time | Memory size when the backup was created |
| Backup file size | Backup file size |
| Cache size | Backup cache size |
| Backup start time | Time when backup creation started |
| Retention expiration | Expiration and deletion time for an automatic backup |
| [⋮] menu | Restore, copy, export, or delete |
View the VPC, availability zone configuration, port, endpoints, topology, attached security groups, and detailed rules. The security group must allow the custom cluster port or default TCP 6379 port.
View the database user name when user authentication is enabled. When transport encryption is enabled, view the expiration date and download the CA file.
See Monitor a cluster.
Monitor a cluster
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, select the cluster.
- View the information on the Monitoring tab.
| Item | Description |
|---|---|
| Monitoring node | Node to monitor; a badge identifies the Primary node |
| Data period | 1 hour (default), 3 hours, 12 hours, 1 day, or 7 days |
| View mode | One, two, or three columns |
| Automatic refresh interval | Off (default), 10 seconds, 30 seconds, 1 minute, or 5 minutes |
| Manual refresh | Refreshes monitoring results |
| Metrics | Engine (Valkey or Redis OSS), CPU, memory, or network |
Monitoring data
| Category | Metric | Description |
|---|---|---|
| Engine | Calls by command (Count/s) | Calls per second by command |
| Engine | Connections (Count) | Maximum and active connections |
| Engine | Cache hit ratio (%) | Cache hit ratio over time |
| Engine | Expired and evicted keys (Count) | Expired and evicted keys over time |
| CPU | Total usage (%) | Total CPU utilization |
| CPU | CPU credits (Count) | Displayed only for nodes that use credits |
| Memory | Total usage (%) | Total memory utilization |
| Memory | Usage by type (Bytes) | Memory usage by type |
| Network | Inbound traffic (Bytes/s) | Bytes received per second |
| Network | Outbound traffic (Bytes/s) | Bytes transmitted per second |
| Network | Received packets (pps) | Packets received per second |
| Network | Transmitted packets (PPS) | Packets transmitted per second |
Configure automatic failover
Valkey automatically enables automatic failover when a cluster has two or more nodes. For Redis OSS, you can select the setting during cluster creation and change it during operation without service interruption. The following procedure applies to Redis OSS.
- Automatic failover is always enabled when cluster mode is enabled and cannot be disabled.
- When cluster mode is disabled, the cluster must be in the
Runningstate and have at least one Replica node to enable automatic failover.
- Go to the KakaoCloud Console > Data Store > MemStore.
- In Redis OSS > Cluster, click [⋮] for the cluster and select Configure automatic failover.
- Select whether to use the feature and click [Save].
- Confirm the setting on the cluster details page.
View security groups
On the Security tab, view the security groups and rules applied to the cluster. The security group must allow the custom cluster port in the 2000–50000 range or the default TCP 6379 port. Click a security group name to manage policies in VPC > Security Group.
Modify security groups
At least one security group must remain attached to the cluster.
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, click [⋮] for the cluster and select Modify security group.
- Select the security groups to apply and review their inbound and outbound rules.
- Click [Modify].
Manage inbound and outbound rules in VPC > Security Group.
Change a parameter group
Select a parameter group compatible with the cluster engine version and cluster mode. Changes to parameter values can affect an active service, and an additional request can fail if a change is already in progress. See Create and manage parameter groups.
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, click [⋮] for the cluster and select Change parameter group.
- Select a parameter group and click [Save].
Change the node type
Changing the node type changes every node in the cluster.
- In a single-node cluster, changing the type restarts the node and causes data loss. Use this feature only for clusters with two or more nodes.
- A switchover can cause a temporary interruption.
- Check workload and memory usage first. High load or insufficient target memory can cause data loss, degraded performance, or a failed request.
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, click [⋮] for the cluster and select Change node type.
- Select a node type and click [Save].
Delete a cluster
Deleted cluster resources are permanently released and cannot be recovered.
- Go to the KakaoCloud Console > Data Store > MemStore.
- In the Cluster menu, click [⋮] for the cluster and select Delete cluster.
- Enter Permanent deletion.
- To retain automatic backups, select
Retain automatic backups.. - Click [Delete].
MemStore user accounts
When user authentication is enabled, authenticate with the ID and password configured during cluster creation. Otherwise, use the passwordless default account.
User authentication is not supported in Redis OSS 5 or earlier.
Command restrictions
For stable service operation, the following Valkey and Redis OSS commands are restricted and return a NOPERM error.
If a node contains many keys, using KEYS, FLUSHALL, or FLUSHDB can stop the node.
- Cluster mode enabled
- Cluster mode disabled
| Command | Description |
|---|---|
| ACL | ACL-related commands |
| BGREWRITEAOF | Rewrite the AOF file |
| BGSAVE | Save in the background |
| CLIENT PAUSE | Pause clients |
| CLUSTER ADDSLOTS | Add slots |
| CLUSTER BUMPEPOCH | Increment the cluster epoch |
| CLUSTER DELSLOTS | Remove slots |
| CLUSTER FAILOVER | Perform cluster failover |
| CLUSTER FLUSHSLOTS | Clear slots |
| CLUSTER FORGET | Disconnect nodes |
| CLUSTER MEET | Connect cluster nodes |
| CLUSTER REPLICATE | Replicate a cluster node |
| CLUSTER RESET | Reset the cluster |
| CLUSTER SAVECONFIG | Save cluster configuration |
| CLUSTER SET-CONFIG-EPOCH | Set the cluster configuration epoch |
| CLUSTER SETSLOT | Change slot settings |
| CONFIG | Configure Redis OSS |
| DEBUG | Debugging commands |
| FAILOVER | Failover command |
| MIGRATE | Migrate data |
| MODULE | Module-related commands |
| PSYNC | Check replication status |
| REPLCONF | Change replication settings |
| REPLICAOF | Configure a Replica |
| RESTORE-ASKING | Request data restoration |
| SAVE | Save data |
| SHUTDOWN | Shut down Redis |
| SLAVEOF | Configure a Replica |
| SORT | Sort data |
| SYNC | Synchronize data |
| Command | Description |
|---|---|
| ACL | ACL-related commands |
| BGREWRITEAOF | Rewrite the AOF file |
| BGSAVE | Save in the background |
| CLIENT PAUSE | Pause clients |
| CLUSTER | Cluster-related commands |
| CONFIG | Configure Redis OSS |
| DEBUG | Debugging commands |
| FAILOVER | Failover command |
| MIGRATE | Migrate data |
| MODULE | Module-related commands |
| PSYNC | Check replication status |
| REPLCONF | Change replication settings |
| REPLICAOF | Configure a Replica |
| RESTORE-ASKING | Request data restoration |
| SAVE | Save data |
| SHUTDOWN | Shut down Redis |
| SLAVEOF | Configure a Replica |
| SORT | Sort data |
| SYNC | Synchronize data |