maxConsecutiveRejects

Stops dropping rejected rows after this many in a row, so a schema change keeps events queued.

This parameter applies to omclickhouse: ClickHouse Output Module.

Name:

maxConsecutiveRejects

Scope:

action

Type:

non-negative integer

Default:

0

Required?:

no

Introduced:

not specified

Description

A row that ClickHouse rejects as bad data is written to errorFile and dropped, so it does not block the rows after it. When every row is rejected, the cause is usually not the data but the table or the template, for example a column whose type changed. Dropping then empties the queue into the error file.

With a value above 0, the action counts final rejections since the last accepted insert. Once the count reaches this value, the next rejected row is not dropped: the action suspends. With action.resumeRetryCount="0" the engine discards the batch; with any other value, while ClickHouse answers its connection check, it retries the batch every action.resumeInterval until ClickHouse accepts it. The rows after it stay queued. The first accepted insert resets the count and rejected rows are dropped again.

0, the default, never stops dropping. Each worker keeps its own count. Size the action queue for the time a schema problem may take to fix.

Action usage

action(type="omclickhouse" maxConsecutiveRejects="100"
       errorFile="/var/log/clickhouse-error.json")

YAML usage

actions:
  - type: omclickhouse
    maxconsecutiverejects: 100
    errorfile: /var/log/clickhouse-error.json

See also

See also omclickhouse: ClickHouse Output Module.


Support: rsyslog Assistant | GitHub Discussions | GitHub Issues: rsyslog source project

Contributing: Source & docs: rsyslog source project

© 2008–2026 Rainer Gerhards and others. Licensed under the Apache License 2.0.