To enable Go runtime metrics collection, start the tracer using the WithRuntimeMetrics option:
tracer.Start(tracer.WithRuntimeMetrics())
View runtime metrics in correlation with your Go services on the Service page in Datadog.
By default, runtime metrics from your application are sent every 10 seconds to the Datadog Agent with DogStatsD. Make sure that DogStatsD is enabled for the Agent. If your Datadog Agent DogStatsD address differs from the default localhost:8125, use the WithDogstatsdAddress option (available starting in 1.18.0) or the environment variables DD_AGENT_HOST and DD_DOGSTATSD_PORT.
If WithDogstatsdAddress is not used, the tracer attempts to determine the address of the statsd service according to the following rules:
/var/run/datadog/dsd.socket and use it if present. IF NOT, continue to #2.DD_AGENT_HOST, and defaults to “localhost”.DD_DOGSTATSD_PORT, and defaults to 8125.If you are running the Agent as a container, ensure that DD_DOGSTATSD_NON_LOCAL_TRAFFIC is set to true, and that port 8125 is open on the Agent. Additionally, for Kubernetes or ECS, follow the guidelines below:
The following metrics are collected by default after enabling Go metrics.
| runtime.go.num_cpu (gauge) | CPUs detected by the runtime. Shown as resource |
| runtime.go.num_goroutine (gauge) | goroutines spawned. Shown as invocation |
| runtime.go.num_cgo_call (gauge) | CGO calls made. Shown as invocation |
| runtime.go.mem_stats.alloc (gauge) | Alloc is bytes of allocated heap objects. Shown as byte |
| runtime.go.mem_stats.total_alloc (gauge) | TotalAlloc is cumulative bytes allocated for heap objects. Shown as byte |
| runtime.go.mem_stats.sys (gauge) | Sys is the total bytes of memory obtained from the OS. Shown as byte |
| runtime.go.mem_stats.lookups (gauge) | Lookups is the number of pointer lookups performed by the Shown as unit |
| runtime.go.mem_stats.mallocs (gauge) | Mallocs is the cumulative count of heap objects allocated. Shown as unit |
| runtime.go.mem_stats.frees (gauge) | Frees is the cumulative count of heap objects freed. Shown as unit |
| runtime.go.mem_stats.heap_alloc (gauge) | HeapAlloc is bytes of allocated heap objects. Shown as byte |
| runtime.go.mem_stats.heap_sys (gauge) | HeapSys is bytes of heap memory obtained from the OS. Shown as byte |
| runtime.go.mem_stats.heap_idle (gauge) | HeapIdle is bytes in idle (unused) spans. Shown as byte |
| runtime.go.mem_stats.heap_inuse (gauge) | HeapInuse is bytes in in-use spans. Shown as byte |
| runtime.go.mem_stats.heap_released (gauge) | HeapReleased is bytes of physical memory returned to the OS. Shown as byte |
| runtime.go.mem_stats.heap_objects (gauge) | HeapObjects is the number of allocated heap objects. Shown as unit |
| runtime.go.mem_stats.stack_inuse (gauge) | StackInuse is bytes in stack spans. Shown as byte |
| runtime.go.mem_stats.stack_sys (gauge) | StackSys is bytes of stack memory obtained from the OS. Shown as byte |
| runtime.go.mem_stats.m_span_inuse (gauge) | MSpanInuse is bytes of allocated mspan structures. Shown as byte |
| runtime.go.mem_stats.m_span_sys (gauge) | MSpanSys is bytes of memory obtained from the OS for mspan structures. Shown as byte |
| runtime.go.mem_stats.m_cache_inuse (gauge) | MCacheInuse is bytes of allocated mcache structures. Shown as byte |
| runtime.go.mem_stats.m_cache_sys (gauge) | MCacheSys is bytes of memory obtained from the OS for Shown as byte |
| runtime.go.mem_stats.buck_hash_sys (gauge) | BuckHashSys is bytes of memory in profiling bucket hash tables. Shown as byte |
| runtime.go.mem_stats.gc_sys (gauge) | GCSys is bytes of memory in garbage collection metadata. Shown as byte |
| runtime.go.mem_stats.other_sys (gauge) | OtherSys is bytes of memory in miscellaneous off-heap Shown as byte |
| runtime.go.mem_stats.next_gc (gauge) | NextGC is the target heap size of the next GC cycle. Shown as byte |
| runtime.go.mem_stats.last_gc (gauge) | LastGC is the time the last garbage collection finished, as nanoseconds since 1970 (the UNIX epoch). Shown as nanosecond |
| runtime.go.mem_stats.pause_total_ns (gauge) | PauseTotalNs is the cumulative nanoseconds in GC Shown as nanosecond |
| runtime.go.mem_stats.num_gc (gauge) | NumGC is the number of completed GC cycles. Shown as unit |
| runtime.go.mem_stats.num_forced_gc (gauge) | NumForcedGC is the number of GC cycles that were forced by the application calling the GC function. Shown as unit |
| runtime.go.mem_stats.gc_cpu_fraction (gauge) | GCCPUFraction is the fraction of this program's available CPU time used by the GC since the program started. Shown as fraction |
| runtime.go.gc_stats.pause_quantiles.min (gauge) | Distribution of GC pause times: minimum values Shown as nanosecond |
| runtime.go.gc_stats.pause_quantiles.25p (gauge) | Distribution of GC pause times: 25th percentile Shown as nanosecond |
| runtime.go.gc_stats.pause_quantiles.75p (gauge) | Distribution of GC pause times: 50th percentile Shown as nanosecond |
| runtime.go.gc_stats.pause_quantiles.95p (gauge) | Distribution of GC pause times: 75th percentile Shown as nanosecond |
| runtime.go.gc_stats.pause_quantiles.max (gauge) | Distribution of GC pause times: maximum values Shown as nanosecond |
Along with displaying these metrics in your APM Service Page, Datadog provides a default Go Runtime Dashboard.
Additional helpful documentation, links, and articles:
