Adding, Updating, and Removing Nodes

Adding a Seq Node

To add a Seq node to the running cluster:

Follow the steps in Configuring a Clustered Seq Node to bring up the node. It will immediately appear in the Data > Cluster screen, if it can join the cluster successfully.
Add the node to the load balancer; see Configuring a Cluster Load Balancer .
In Settings > General, update the cluster's "Target replica count" setting.

🚧
The order of steps is important: the node must not receive ingestion traffic until it has joined the cluster. If ingestion traffic (from the load balancer) does reach the node prematurely, it will refuse to join the cluster so that data it has ingested is not lost.

Updating a Seq Node

To update a node in a Seq cluster to a new Seq version, or to perform other maintenance on the node:

Check that the cluster is healthy; see Data > Cluster, the /health/cluster endpoint, or the seqcli cluster health command.
Stop the node using seq service stop (Windows) or docker stop (Docker). The node can be hard-terminated if necessary, but if the node is leading the cluster, this may cause a short period of unavailability (by default, ~15s) while a new cluster leader is elected.
Perform the required maintenance.
Start the node using seq service start (Windows) or docker start (Docker).
Wait for the cluster to become healthy again; seqcli cluster health --wait-until-healthy is helpful for this purpose in automation scenarios.

Removing a Seq Node

To remove a node from a Seq cluster, assuming it is not the last node:

Check that the cluster is healthy; see Data > Cluster, the /health/cluster endpoint, or the seqcli cluster health command.
Stop the node using seq service stop(Windows) ordocker stop\ (Docker).
Terminate/decommission the node.
In Settings > General, reduce the cluster's "Target replica count".

❗️
Do not remove the last healthy node in a Seq cluster. Application log and trace data is stored locally on the cluster's Seq nodes, so removing the last replica will empty the cluster of data.