Pulling data from a statsd server

[humartinez]

Hi,
I need to pull data from a statsd server,whats the best approach to do this? Im trying to import the statsd metrics from a spark jobs using metricbeats but metricbeat is getting the following error

...
spark-5a9e0fdbde2e44b38a075a6bdbb67ecb_d
river_LiveListenerBus_queue_shared_numDroppedEvents":{"count":0},"spark-5a9e0fdbde2e44b38a075a6bdbb67ecb_driver_LiveListenerBus_queue_shared_size":{"value":0.000000}}}, Private:interface {}(nil), TimeSeries:true}, Flags:0x0, Cache:publisher.EventCache{m:common.MapStr(ni
l)}} (status=400): {"type":"mapper_parsing_exception","reason":"failed to parse","caused_by":{"type":"illegal_argument_exception","reason":"Limit of total fields [10000] has been exceeded while adding new fields [3621]"}}

Is there a workarround to this? Is there a better way to ingests these metrics?

Thanks in advance

[warkolm (Mark Walkom)]

TLDR this is managed by Mapping limit settings | Elasticsearch Guide [8.11] | Elastic, and it looks like you have already increased it from the default of 1000. Increasing it more isn't a great idea as it causes mapping explosions and inefficiencies.

@humartinez

Limit of total fields [10000] has been exceeded while adding new fields

What other data are you pulling in from Metricbeat?
Can you put data from different sources into different indices?

[humartinez]

Im trying to pull data from some sparks jobs via statsd. I think that the whole issue is about metrics mappings, without it the ELK stack is trying to index the metrics with the sparkIds as a field thats why the number of records exceds the limits.

Regarding to mappings how do i translate a graphite map like this

- match: '*.*.LiveListenerBus.listenerProcessingTime.org.apache.spark.sql.execution.ui.*.*'
  name: Live_Listener_Bus_ExecutionUI
  labels:
    application: $1
    executor_id: $2
    type: $3
    qty: $4

Currently Im starting to use a graphite exporter to get de metrics from spark, and translate them to prometheus metrics labels. But I would really like to try Elastic as a metrics collector

Thanks