Skip to main content

Adding Logs

Typically the Embrace SDK uploads data at the end of a session. However, some situations might require instant feedback, such as hunting an especially difficult bug, troubleshooting on behalf of high-value users, or monitoring a new version rollout.

You can leverage the log message API for this.

Using the Log Message API​

The Log Message API enables you to log a message instantly.

Depending on your requirements, you can choose from three severity types: logInfo, logWarning, and logError.

To use this API, you need to pass the following arguments to the method:

  1. The message: The string is the message itself. Try to make this short yet informative.
  2. Properties: A map of key-value pairs that enables you to categorize and filter log messages.

In the case of logError, you may also send an Exception to be shown on the dashboard.

Here is an example of how to use the Log Message API for errors:

Embrace.getInstance().logError("Loading not finished in time.")

Or you can call logMessage if you want to add properties to your logs.

val props = mutableMapOf<String, Any>()
props["propertyA"] = "valueA"
props["propertyB"] = "valueB"
Embrace.getInstance().logMessage("Loading not finished in time.", Severity.ERROR, props)

Log Handled Exception​

If there is a need to log an exception, but the severity level is something other than an error, the logException method can be used.

val props = mutableMapOf<String, Any>()
props["propertyA"] = "valueA"
props["propertyB"] = "valueB"

try {
    val exception = NullPointerException("this is my handled exception")
    throw exception
} catch (e: Exception) {
    Embrace.getInstance().logException(e, Severity.WARNING, props)
}

LogType could be ERROR, WARNING or INFO.

Adjusting Severity

You can also adjust the severity of the log by either calling the logWarning or logInfo methods.

Embrace.getInstance().logWarning("User attempted expired credit card")
Embrace.getInstance().logInfo("User has entered checkout flow")
Limits on Log Messages
  • The maximum length for a log message is 128 characters. Messages are truncated if they exceed the limit.
  • Properties are limited to 10 per log
  • Property keys have a limit of 128 characters.
  • Property values have a limit of 256 characters.

Log Batching​

In some cases, logs will be batch and sent at once, to reduce the device and network overhead. We batch logs according to the following criteria:

  • MAX_LOGS_PER_BATCH = 50. Once we reach 50 logs, we send the batch (all in a single request)
  • MAX_BATCH_TIME = 5000 (in milliseconds). We send the batch 5 seconds after the first log in the batch was received.
  • MAX_INACTIVITY_TIME = 2000 (in milliseconds). If no new log appears after 2 seconds from the last log, we send the batch.

Being Alerted on Logs​

Once you start using our alerting feature you can also configure how these are handled on the backend. Using the Embrace Dashboard, you can configure email alerts to be sent to your team when certain thresholds are met with log events. For example, let's say you have a steady rate of 1% for a given log event. You could set that as a threshold and receive an email if the rate rises beyond that in a sustained way.

Export your telemetry​

A LogRecordExporter can be easily injected, to directly export your data to any OpenTelemetry Collector.

Local testing​

Injecting a SystemOutLogRecordExporter will allow you to see your telemetry in the logcat.

2024-03-05 14:15:15.342 29672-29756 System.out     io.embrace.mysampleapp          I  1970-01-01T00:00:00Z INFO 'Default log'

Adding a LogRecordExporter for a custom OTel Collector​

You can send your data to a custom (OTel Collector)

 //grpc through an otel collector in a local docker image
val customDockerLogRecordExporter = OtlpGrpcLogRecordExporter.builder()
    .setEndpoint("http://10.0.2.2:4317")
    .build()

Exporting data to Grafana Cloud​

Embrace Logs can be exported to Grafana Cloud using an OTel Collector.

//... or directly to grafana cloud
val grafanaCloudLogRecordExporter = OtlpHttpLogRecordExporter.builder()
    .setEndpoint("https://myinstance.grafana.net/otlp/v1/traces")
    .addHeader("Authorization", "YourToken")
    .build()
info

Every exporter should be added before starting the SDK

Embrace.getInstance().addLogRecordExporter(SystemOutLogRecordExporter.create())
Embrace.getInstance().addLogRecordExporter(customDockerLogRecordExporter)
Embrace.getInstance().addLogRecordExporter(grafanaCloudLogRecordExporter)

Embrace.getInstance().start(this)

Best Practices​

Logging a message using the Log Message API makes a network request immediately. Sending too many logs can easily impact application performance or battery life.

info

For more tips on making the most of the Log Message API, checkout the Best Practices.