[
https://issues.apache.org/jira/browse/FLINK-21747?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17300246#comment-17300246
]
Chesnay Schepler commented on FLINK-21747:
------------------------------------------
Apparently the
[64kb|https://docs.influxdata.com/influxdb/v1.7/write_protocols/line_protocol_reference/#data-types]
that InfluxDB imposes applies to the entire set of tags.
You should be able to work around the issue by excluding these specific
variables like this:
{code}
metrics.reporter.<name>.scope.variables.excludes: task_name;operator_name
{code}
> Encounter an exception that contains "max key length exceeded ..." when
> reporting metrics to influxdb
> ------------------------------------------------------------------------------------------------------
>
> Key: FLINK-21747
> URL: https://issues.apache.org/jira/browse/FLINK-21747
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Metrics
> Affects Versions: 1.10.0
> Reporter: tim yu
> Priority: Major
>
> I run a stream job with insert statement whose size is too long, it report
> metrics to influxdb. I find many influxdb exceptions that contains "max key
> length exceeded ..." in the log file of job manager . The job could not write
> the metrics to the influxdb, because "task_name" and "operator_name" is too
> long.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
