| Name | Description | Default value |
|---|---|---|
| thread_pools | Thread pools options. | – |
| thread_pools.redis_thread_pool_size | Thread count to serve Redis requests. | – |
| thread_pools.sentinel_thread_pool_size | Thread count to serve sentinel requests. | – |
| groups.[] | Redis cluster to work with excluding subscribers. | – |
| groups.[].config_name | Key name in secdist with options for this cluster. | – |
| groups.[].db | Name to refer to the cluster in components::Redis::GetClient(). | – |
| groups.[].sharding_strategy | One of RedisStandalone, RedisCluster, KeyShardCrc32, KeyShardTaximeterCrc32 or KeyShardGpsStorageDriver. Possible values: RedisCluster, KeyShardCrc32, KeyShardTaximeterCrc32, KeyShardGpsStorageDriver, RedisStandalone. | KeyShardTaximeterCrc32 |
| groups.[].allow_reads_from_master | Allows read requests from master instance. | false |
| groups.[].topology_update_method | Topology update method for cluster databases. cluster_slots - use CLUSTER SLOTS command (default). cluster_shards - use CLUSTER SHARDS command. Possible values: cluster_slots, cluster_shards. | cluster_slots |
| groups.[].required | Required connection mode for the client. Possible values: no_wait, master, slave, master_or_slave, master_and_slave. | – |
| groups.[].max_disconnect_time | The maximum allowed time in seconds for a database host to be disconnected before it is considered unhealthy. It depends on the cluster-node-timeout setting in the Redis/Valkey config and should be greater than 2x that value plus some additional time. | 35 |
| groups.[].max_failed_shards | Maximum number or percent of failed shards allowed in GetComponentHealth check. 0 means all shards must be ready (default behavior). This only affects health checks, not WaitConnectedOnce. | – |
| groups.[].max_failed_shards.amount | Maximum number of failed shards allowed. | 0 |
| groups.[].max_failed_shards.percent | Maximum percent of failed shards allowed. | 0 |
| metrics_level | Set metrics detail level. Possible values: cluster, shard, instance. | instance |
| subscribe_groups.[] | Redis cluster to work with in subscribe mode. | – |
| subscribe_groups.[].config_name | Key name in secdist with options for this cluster. | – |
| subscribe_groups.[].db | Name to refer to the cluster in components::Redis::GetSubscribeClient(). | – |
| subscribe_groups.[].sharding_strategy | Either RedisCluster or KeyShardTaximeterCrc32. Possible values: RedisCluster, KeyShardTaximeterCrc32, RedisStandalone. | KeyShardTaximeterCrc32 |
| subscribe_groups.[].allow_reads_from_master | Allows subscriptions to master instance to distribute load. | false |
| subscribe_groups.[].topology_update_method | Topology update method for cluster databases. cluster_slots - use CLUSTER SLOTS command (default). cluster_shards - use CLUSTER SHARDS command. Possible values: cluster_slots, cluster_shards. | cluster_slots |
| subscribe_groups.[].per_channel_statistics | Enables statistics for each redis pub/sub channel. | true |
| subscribe_groups.[].required | Required connection mode for the subscription client. Possible values: no_wait, master, slave, master_or_slave, master_and_slave. | – |
| subscribe_groups.[].max_disconnect_time | The maximum allowed time in seconds for a database host to be disconnected before it is considered unhealthy. It depends on the cluster-node-timeout setting in the Redis/Valkey config and should be greater than 2x that value plus some additional time. | 35 |
| subscribe_groups.[].max_failed_shards | Maximum number or percent of failed shards allowed in GetComponentHealth check. 0 means all shards must be ready (default behavior). This only affects health checks, not WaitConnectedOnce. | – |
| subscribe_groups.[].max_failed_shards.amount | Maximum number of failed shards allowed. | 0 |
| subscribe_groups.[].max_failed_shards.percent | Maximum percent of failed shards allowed. | 0 |
