Skip to main content

Prometheus Remote Write

To send your Prometheus application metrics to a Logz.io Infrastructure Monitoring account, use remote write to connect to Logz.io as the endpoint. Your data is formatted as JSON documents by the Logz.io listener.

note

Multiple server environments: If you have multiple Prometheus server instances, you'll have to add Logz.io as an endpoint for each instance.

Configuring Remote Write to Logz.io

Get your Logz.io Infrastructure Monitoring account information

Within Logz.io, look up the Listener host for your region (URL) and the Logz.io Metrics Account token.

Add a remote_write url

Configure your Prometheus yaml file or use a Helm chart:

To configure a Prometheus yaml file

Add the following parameters to your Prometheus yaml file:

Environment variableDescriptionRequired/Default
external_labelsParameters to tag the metrics from this specific Prometheus server.
p8s_logzio_nameUse the value of the parameter p8s_logzio_name to identify from which Prometheus environment the metrics are arriving to Logz.io. Replace the <labelvalue> placeholder with a label that will be added to all the metrics that are sent from this specific Prometheus server.
remote_writeThe remote write section configuration sets Logz.io as the endpoint for your Prometheus metrics data. Place this section at the same indentation level as the global section.
urlThe Logz.io Listener URL for your region, configured to use port 8052 for http traffic, or port 8053 for https traffic. For more details, see the Prometheus configuration file remote write reference. Required
bearer_tokenThe Logz.io Prometheus Metrics account token.Required
global:
external_labels:
p8s_logzio_name: <labelvalue>
remote_write:
- url: https://<<LISTENER-HOST>>:8053
bearer_token: <<PROMETHEUS-METRICS-SHIPPING-TOKEN>>
remote_timeout: 30s
queue_config:
batch_send_deadline: 5s #default = 5s
max_shards: 10 #default = 1000
min_shards: 1
max_samples_per_send: 500 #default = 100
capacity: 10000 #default = 500

To configure a Helm chart

For kube-prometheus-stack Helm chart users:

Edit your chart values.yaml file in the following sections:

  • remote write:
remoteWrite:
- url: https://<<LISTENER-HOST>>:8053 # The Logz.io Listener URL for your region, configured to use port **8052** for http traffic, or port **8053** for https traffic.
bearerToken: <<PROMETHEUS-METRICS-SHIPPING-TOKEN>> # The Logz.io Prometheus metrics account token
remoteTimeout: 30s
queueConfig:
batchSendDeadline: 5s #default = 5s
maxShards: 10 #default = 1000
maxSamplesPerSend: 500 #default = 100
capacity: 10000 #default = 500
  • externalLabels:
externalLabels:
p8s_logzio_name: <labelvalue>
Verify the remote_write configuration
  • Run a query: If you are scraping Prometheus metrics, you can check that the remote_write configuration is working properly by doing one of the following and verifying that the result is greater than zero (n > 0) for the url:

    • Run a query on your local Prometheus interface for the metric prometheus_remote_storage_samples_in_total.

    • Check for the metric in the /metrics endpoint on your Prometheus server.

  • Check via Metrics Explore: To verify that metrics are arriving to Logz,io:

    1. Open Metrics Explore via the compass icon in the left menu bar.

    2. Examine the Metrics dropdown list below the Explore heading in the upper left of the pane. An empty list or the text no metrics indicates that the remote write configuration is not working properly. Verify Prometheus metrics

Open Metrics Explore

Once you've verified that your data is available in Logz.io, explore your Prometheus metrics.

After your metrics are flowing, import your existing Prometheus and Grafana dashboards to Logz.io Infrastructure Monitoring as JSON files.

Performance tips

  • Reduce tagging: By default, all the metrics from your Prometheus server(s) are sent to Logz.io. To drop or send specific metrics, add Prometheus labeling before enabling the remote write, or as part of the remote write configuration. Learn more about Prometheus relabeling tricks here.
  • Metrics metadata dashboards: If you have both Prometheus & Grafana, you can activate a dashboard as part of the remote write configuration that will show you the queue size and how many metrics you're sending. If your queue size increases, it might be necessary to open an additional channel. (currently in development)

  • Tune the remote write process: Learn more about Prometheus remote write tuning here.