[Enhancement Request] Flow: OTLP logs exporter to Loki

[cyrille-leclerc]

Enhancement Request

In order to fully leverage the Grafana Agent OTLP receiver (otelcol.receiver.otlp), I want to not only send my metrics to Mimir and traces to Tempo but I also want to send my logs to Loki.

To do so, I imagine I would define a Grafana Agent Flow pipeline with an otelcol.exporter. for Loki similar to the otelcol.exporter.prometheus. I would expect consistency of the default setup of the Loki exporter with the Grafana OTLP Gateway.

I am aware of the impedance challenges between OTel Logs and Loki, typically on the mapping of the structured field of OTel logs, but

• Those challenges are IMO not very different from the mapping challenges we have between OTel metrics and Mimir
• I would expect to solve these challenges looking at the user workflows like the workflow to enable slicing and dicing on logs

See Also

• https://github.com/grafana/agent/issues/2213

[cyrille-leclerc]

Here is an experiment on the UX introducing otelcol.exporter.loki, see PROPOSED NEW PIPELINE STEP blocks:

otelcol.receiver.otlp "default" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/otelcol.receiver.otlp/
    grpc {
        endpoint = "0.0.0.0:4318"
    }

    output {
        metrics = [otelcol.processor.batch.default.input]
        logs    = [otelcol.processor.batch.default.input]
        traces  = [otelcol.processor.batch.default.input]
    }
}

otelcol.processor.batch "default" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/otelcol.processor.batch/
    output {
        metrics = [otelcol.processor.memory_limiter.default.input]
        logs    = [otelcol.processor.memory_limiter.default.input]
        traces  = [otelcol.processor.memory_limiter.default.input]
    }
}

otelcol.processor.memory_limiter "default" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/otelcol.processor.memory_limiter/
    check_interval = "1s"

    limit = "50MiB" // alternatively, set `limit_percentage` and `spike_limit_percentage`

    output {
        metrics = [otelcol.exporter.prometheus.default.input]
        // logs    = [otelcol.exporter.loki.default.input] // PROPOSED NEW PIPELINE STEP
        traces = [otelcol.exporter.otlp.grafana_cloud_tempo.input]
    }
}

// PROPOSED NEW PIPELINE STEP
//otelcol.exporter.loki "default" {
//  forward_to = [loki.write.default.receiver]
//}

otelcol.exporter.prometheus "default" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/otelcol.exporter.prometheus/
    forward_to = [prometheus.remote_write.default.receiver]
}

otelcol.exporter.otlp "grafana_cloud_tempo" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/otelcol.exporter.otlp/
    client {
        endpoint = env("TEMPO_ENDPOINT")
        auth     = otelcol.auth.basic.grafana_cloud_tempo.handler
    }
}

loki.write "default" {
    // https://grafana.com/docs/agent/latest/flow/reference/components/loki.write/
    endpoint {
        url = env("LOKI_URL")
        http_client_config {
            basic_auth {
                username = env("LOKI_USERNAME")
                password = env("LOKI_PASSWORD")
            }
        }
    }
}

...

See full flow configuration here https://github.com/cyrille-leclerc/my-shopping-cart/blob/d4556f46f93203f0fd055ecaea005809391b2a9d/grafana/agent-config.flow

[tpaschalis]

Thanks for the interest! We're already tracking the inclusion of such OTLP <-> Loki adapters in #2501, apologies if it was hidden away as a separate ticket.

The adapters would work exactly as you envisioned here, glad to see that it makes sense for someone already using Flow!

[rfratto]

Thanks for the feature request! As @tpaschalis mentioned, this is a dupe of #2501, #2665, and #2666, so I'm going to close this in favor of those.