Monitoring & Metrics

All good micro-services require a standard way of measuring & monitoring system performance, providing health checks and so on - Respite comes with first-class support for this with @coda`s Metrics library.

Health Checks

Health checks are just that - they confirm if a service is available, healthy and ready to service clients or not.You can find them at /metrics/health

$ curl -X GET -H"content-type: application/json" "http://localhost:8080/metrics/health"

Creating custom Health Checks

By default, RestController objects contain a default health check that queries "/"(all) which is executed when the metrics, however if you would like to create a custom monitor simply wrap the check in a healthCheck call as per below:

   * Create default health check on REST controller - confirm CRUD pipelines are active
   * (what exactly does 'active' mean?)
  healthCheck("doSomething", s"doSomething method in $metricBaseName") {
    def check(): Boolean = {
      // Default to failed health check

      try {
        // Perform some operation
        // ...

        // True will mark health check up
      } catch {
        // False will fail the health check

    // See

Note: the check() function here is necessary due to an issue with the underlying framework. The variable $metricBaseName is available and refers to the parents class name, and is useful for identifying the originating class MetricsSupport was mixed in to.

Runtime Metrics

Respite by default will instrument & measure calls to all Routes on a given Controller using things called Timers. For CRUD services (aka RestControllers), it will give sensible names to each timer such as 'list', 'delete' and 'get'. For non-RestControllers, it will use the path provided in the URL as the name of the Timer:

Runtime metrics for instrumented Controllers are available at /metrics/:

$ curl -X GET -H"content-type: application/json" "http://localhost:8080/metrics/"
    "version": "3.0.0",
    "gauges": {},
    "counters": {},
    "histograms": {},
    "meters": {},
    "timers": {
        "": {
            "count": 10000,
            "max": 0.008023,
            "mean": 0.00009569163424124514,
            "min": 0.000043,
            "p50": 0.00006,
            "p75": 0.000063,
            "p95": 0.000087,
            "p98": 0.00012203999999999952,
            "p99": 0.00029594000000000056,
            "p999": 0.00802068,
            "stddev": 0.0004458142701117825,
            "m15_rate": 1630.4,
            "m1_rate": 1630.4,
            "m5_rate": 1630.4,
            "mean_rate": 1596.013541217289,
            "duration_units": "seconds",
            "rate_units": "calls/second"

For further reading on Metrics library, visit the documentation on the DropWizard website.