[influxdb] InfluxDB OOM issue - How to set shard duration with cache-max-memory-size ？

[qli via InfluxData]

I deployed single InfluxDB (1.4.2) node with multiple DBs, we used telegraf to collect system metrics into different DBs (isolate data base on security policy)

recently days (after running 2 months), always met InfluxDB OOM issues, according Influx official docs, I thinks there is some issues with shard duration and series cardinality

Retention Policy is 90 days for each DB, and I found each shard or shard group size is vary small (<100MB)

1. Is cache-max-memory-size max limit for per shard ?
if so the default value (1G) is not right size for my case, should I reduce cache-max-memory-size to smaller value and increase shard duration from 1 day to higher value like 7 days ?


2.series cardinality with hostname tag
Our servers is dynamically created and destroyed, so there will up to 10 thousands servers (even more) hostname as one of tags, so it will increase series cardinality (at least x * 10000), how can i handle it ?
we need hostname to group by or filter metrics sometimes, but most of time we query metrics coarse granularity with other tags. Looks like it is not a good idea put hostname into fields set.

is there better solution for the hostname tag ?
--
*This email may contain or reference confidential information and is
intended only for the individual to whom it is addressed. Please refrain
from distributing, disclosing or copying this email and the information
contained within unless you are the intended recipient. If you received
this email in error, please notify us at ***@appannie.com
<***@appannie.com>** immediately and remove it from your system.*