Replication and Keeper health

Answers "are the replicas connected, in sync, and draining their queues?" from system.zookeeper_connection, system.replicas, system.replication_queue, system.replicated_fetches, system.events and system.text_log. Run altinity-expert-clickhouse-connection first if the connection mode, cluster and time window are not yet established.

Query packs

• triage.sql — 2 checks: Keeper/ZooKeeper session status per host (replication-triage-01, @requires keeper), and the graded replication overview from system.replicas (replication-triage-02). Start here.
• queue.sql — 2 checks: graded queue size per table and host (replication-queue-01), and the individual queue tasks carrying an exception or a postpone reason (replication-queue-02).
• fetches.sql — 2 checks: fetches in flight with elapsed time and progress (replication-fetches-01), and recent DownloadPart activity (replication-fetches-02, needs system.part_log).
• keeper.sql — 2 checks: average Keeper round-trip latency per host derived from ZooKeeperWaitMicroseconds over ZooKeeperTransactions (replication-keeper-01), and Keeper errors and warnings from the last 24 hours (replication-keeper-02, needs system.text_log).

How to run the query packs