Logging & Monitoring

Cloud Run sends everything your service logs to Cloud Logging, with no setup. Declare the checks and alerts that watch it next to the service itself.

Log from your code

Effect.log goes to Cloud Logging:

yield* Effect.log(`order ${id} confirmed`);
yield* Effect.logWarning(`retrying ${id}`);

Check that the service is up

An uptime check probes a URL from several regions:

Effect.gen(function* () {
  const api = yield* Api;

  const check = yield* GCP.Monitoring.UptimeCheckConfig("ApiUp", {
    httpCheck: { path: "/", useSsl: true, validateSsl: true },
    monitoredResource: {
      type: "uptime_url",
      labels: { host: Output.map(api.uri, (uri) => new URL(uri!).host) },
    },
    period: "300s",
  });
}),

Get notified

Create a channel to send alerts to:

const oncall = yield* GCP.Monitoring.NotificationChannel("Oncall", {
  type: "email",
  labels: { email_address: "oncall@example.com" },
});

Alert on a metric

An alert policy opens an incident when a condition holds. This one fires when the service answers with 5xx errors:

yield* GCP.Monitoring.AlertPolicy("ApiErrors", {
  combiner: "OR",
  conditions: [
    {
      displayName: "5xx responses",
      conditionThreshold: {
        filter:
          'resource.type = "cloud_run_revision" AND metric.type = "run.googleapis.com/request_count" AND metric.labels.response_code_class = "5xx"',
        comparison: "COMPARISON_GT",
        thresholdValue: 5,
        duration: "60s",
        aggregations: [{ alignmentPeriod: "60s", perSeriesAligner: "ALIGN_RATE" }],
      },
    },
  ],
  notificationChannels: [oncall.name],
});

The condition has to hold for its whole duration before the policy opens an incident and notifies the channel. When the errors stop, the incident closes.

Count log lines

A logs-based metric turns matching log entries into a metric you can chart and alert on:

const errors = yield* GCP.Logging.Metric("Errors", {
  filter: 'resource.type = "cloud_run_revision" AND severity >= ERROR',
  description: "error log entries",
});

Route logs elsewhere

A sink copies matching entries to another destination, such as a log bucket or BigQuery:

yield* GCP.Logging.Sink("ErrorArchive", {
  destination:
    "logging.googleapis.com/projects/my-project/locations/global/buckets/_Default",
  filter: "severity >= ERROR",
});

Where next

• Observability — tracing Effect programs.
• Monitoring reference and Logging reference — every prop.