[PROF-10275] Enable crashtracking telemetry by default when profiler …

…is enabled

**What does this PR do?**

This PR enables the crashtracking telemetry feature by default when the
profiler is enabled.

This feature can be disabled using the
`DD_PROFILING_EXPERIMENTAL_CRASH_TRACKING_ENABLED` environment variable
or the `c.profiling.advanced.experimental_crash_tracking_enabled`
setting via code.

This setting is not very well-named, as the feature is no longer
experimental. But as we're actually considering separating crashtracking
from profiling, I decided to leave the current setting in place until we
can decide where this is going to live, since that's going to influence
where the new setting without "experimental" on its name should land.

**Motivation:**

Crashtracking allows us to identify issues that may otherwise go
unnoticed. We've been deploying it on other Datadog libraries with great
success so far.

**Additional Notes:**

We've had configurations running with crashtracking enabled on our usual
validation environments + almost all (1 remaining!) internal Ruby apps
also has been running with this feature.

We've not seen any noticeable impact for enabling this feature.

In fact, now that it's going to be the default, I'll go ahead and remove
the separate configurations we had for testing it.

**How to test the change?**

The following snippet starts a Ruby instance with profiling, and crashes
it immediately:

```bash
DD_PROFILING_ENABLED=true bundle exec ddprofrb exec ruby -e 'Process.kill("SEGV", Process.pid)'
```

You should see an error report submitted to the agent.

.gitlab/benchmarks.yml

@@ -109,13 +109,6 @@ only-profiling-heap:
     DD_PROFILING_ALLOCATION_ENABLED: "true"
     DD_PROFILING_EXPERIMENTAL_HEAP_ENABLED: "true"
 
-only-profiling-crashtracking:
-  extends: .benchmarks
-  variables:
-    DD_BENCHMARKS_CONFIGURATION: only-profiling
-    DD_PROFILING_ENABLED: "true"
-    DD_PROFILING_EXPERIMENTAL_CRASH_TRACKING_ENABLED: "true"
-
 profiling-and-tracing:
   extends: .benchmarks
   variables:

lib/datadog/core/configuration/settings.rb

@@ -453,12 +453,15 @@ def initialize(*_)
 
             # Enables reporting of information when the Ruby VM crashes.
             #
+            # This feature is no longer experimental, and we plan to deprecate this setting and replace it with a
+            # properly-named one soon.
+            #
             # @default `DD_PROFILING_EXPERIMENTAL_CRASH_TRACKING_ENABLED` environment variable as a boolean,
-            # otherwise `false`
+            # otherwise `true`
             option :experimental_crash_tracking_enabled do |o|
               o.type :bool
               o.env 'DD_PROFILING_EXPERIMENTAL_CRASH_TRACKING_ENABLED'
-              o.default false
+              o.default true
             end
           end
 

lib/datadog/profiling/component.rb

@@ -126,7 +126,7 @@ def self.build_profiler_component(settings:, agent_settings:, optional_tracer:)
         # and thus can't really provide a valid configuration to talk to a Datadog agent. Thus, in this situation,
         # we can't use the crashtracker, even if enabled.
         unless transport.respond_to?(:exporter_configuration)
-          Datadog.logger.warn(
+          Datadog.logger.debug(
             'Cannot enable profiling crash tracking as a custom settings.profiling.exporter.transport is configured'
           )
           return

spec/datadog/core/configuration/settings_spec.rb

@@ -828,7 +828,7 @@
           context 'is not defined' do
             let(:environment) { nil }
 
-            it { is_expected.to be false }
+            it { is_expected.to be true }
           end
 
           [true, false].each do |value|
@@ -843,10 +843,10 @@
 
       describe '#experimental_crash_tracking_enabled=' do
         it 'updates the #experimental_crash_tracking_enabled setting' do
-          expect { settings.profiling.advanced.experimental_crash_tracking_enabled = true }
+          expect { settings.profiling.advanced.experimental_crash_tracking_enabled = false }
             .to change { settings.profiling.advanced.experimental_crash_tracking_enabled }
-            .from(false)
-            .to(true)
+            .from(true)
+            .to(false)
         end
       end
 

spec/datadog/profiling/component_spec.rb

@@ -547,8 +547,6 @@
       end
 
       context 'when crash tracking is enabled' do
-        before { settings.profiling.advanced.experimental_crash_tracking_enabled = true }
-
         it 'initializes the crash tracker' do
           expect(Datadog::Profiling::Crashtracker).to receive(:new).with(
             exporter_configuration: array_including(:agent),
@@ -564,11 +562,11 @@
 
           before do
             settings.profiling.exporter.transport = custom_transport
-            allow(Datadog.logger).to receive(:warn)
+            allow(Datadog.logger).to receive(:debug)
           end
 
-          it 'warns that crash tracking will not be enabled' do
-            expect(Datadog.logger).to receive(:warn).with(/Cannot enable profiling crash tracking/)
+          it 'debug logs that crash tracking will not be enabled' do
+            expect(Datadog.logger).to receive(:debug).with(/Cannot enable profiling crash tracking/)
 
             build_profiler_component
           end