Smart client handoffs

Learn how to avoid disruptions during Redis server maintenance.

Smart client handoffs (SCH) is a feature of Redis Cloud and Redis Software servers that lets them actively notify clients about planned server maintenance shortly before it happens. This lets a client reconnect or otherwise respond gracefully without significant interruptions in service.

SCH is primarily useful when server software or hardware upgrades are required. It provides two main features to help the client avoid disruptions in service during the maintenance period:

  • Relaxed timeouts: Upgrades tend to impact the general performance of the server. Advance notification of the upgrade lets a client adjust its command timeouts to take this into account and avoid aborting commands too soon.
  • Pre-handoffs: Upgrades also involve migrating Redis shards to new nodes, which inevitably disconnects clients from existing nodes. However, with some advance warning of the disconnection, a client can buffer commands, connect to a new node, and then resume the buffered commands without aborting any of them. As a result, users see no disruption in service. These transparent reconnections to new endpoints are known as pre-handoffs.
Note:
SCH does not work with blocking connections. These include connections used for blocking operations like BLPOP and also pub/sub subscriptions. All non-blocking operations are safe to use with SCH.

SCH support in Redis client libraries

SCH is enabled automatically on the client side during connection if you select the RESP3 protocol, which is a requirement for SCH. However, you can configure some parameters, such as the timeouts to use during maintenance. See the pages linked below to learn how to configure SCH for:

SCH support in Redis server products

Redis Cloud

SCH is fully supported and enabled by default on Redis Cloud, except when you are using one of the following options:

These services don't currently allow for pre-handoffs, but you still get the benefit of relaxed timeouts during database version upgrades. All other configurations have full support for both relaxed timeouts and pre-handoffs.

Redis Software

You must enable SCH explicitly on self-managed Redis Software servers by using the v1/cluster REST API request to set the client_maint_notifications option to true. The example below shows how to do this using the curl command line utility:

curl -k -X PUT -H "accept: application/json" \
    -H "content-type: application/json" \
    -u "[email protected]:test123" \
    -d '{ "client_maint_notifications": true }' \
    https://localhost:9443/v1/cluster

Redis Software uses relaxed timeouts to avoid command failures during database version upgrades. The support for pre-handoffs depends on the specific upgrade method you use, as detailed in the table below.

Upgrade method SCH support Expected behavior
Rolling upgrade Full New nodes and old ones removed sequentially. SCH pre-handoffs and relaxed timeouts greatly reduce disruptions during the upgrade.
In-place upgrade Partial Relaxed timeouts reduce errors but there are no pre-handoffs. Disconnections occur when processes are replaced during the upgrade, so clients should rely on auto-reconnect, which will cause brief lapses in service.
Maintenance mode Full SCH is fully supported during hardware or OS patching operations. Pre-handoffs and relaxed timeouts minimize application impact.

Redis Enterprise for Kubernetes

SCH is not currently supported for Kubernetes clusters.

Redis Open Source

SCH is not currently supported for Redis Open Source.

RATE THIS PAGE
Back to top ↑