Service_name is always unknown_service

Grafana Loki
1

Hi,

I setup a loki instance and started ingesting the logs using kube logging operator.

But it seems the service_name is always unknown_service.

I tried sending different fields like service_name, service.name and service, but nothing works it still shows the same unknown_service.

i restarted the loki cluster also but it didnt work.

example value of my service “monitoring______grafana__prometheus“

is this a known issue? or am i missing something?

your help is deeply appreciated,

thanks

2

Try to send desired service.name as resource (not log line) attribute.

3

got it, to set service as resource, i added the below in loki configuration:

labels:

  service: $.service

This resolved my issue and started grouping the logs properly.

4

@vipinjn24 @jangaraj , we are facing same unknown_service issue, Plugin used is fluent-plugin-grafana-loki. We tried below config in Loki chart and the other config on fluentd side, please advise:

Loki-chart-config:

limits_config:
discover_service_name:
- index

Fluentd-side config:

 <label>

            index $.index

            service_name $.index

            service $.index

          </label>

full fluentd loki-config:

      <label @REWRITTEN>

        <match \*\*>

          @type loki

          url "#{ENV\['LOKI_URL'\]}"

          ca_cert /secure/tls-certs/ca.crt

          line_format json

      \# FIX 1: Provide ONLY 'index' and 'valli'. 

      \# Remove 'service_name' and 'service' so Loki's chart can discover 'index' automatically.

          <label>

            valli $.index

            index $.index

            service_name $.index

            service $.index

          </label>

      \# FIX 2: Do NOT include 'remove_keys index'. 

      \# If you remove it here, Fluentd's buffer chunking breaks during evaluation.



      \# FIX 3: Keep $.index as the primary memory chunk routing key.

      <buffer $.index>

        @type memory

        flush_interval 1s

        retry_type periodic

        retry_forever false

        retry_wait 10s

        retry_max_times 2

        chunk_limit_size 8m

        total_limit_size 256MB

        overflow_action drop_oldest_chunk

        flush_thread_count 16

        queued_chunks_limit_size 32

        flush_at_shutdown true

      </buffer>

    </match>

  </label>