Enable service default sampling param

[defool]

Signed-off-by: defool defool@foxmail.com

Which problem is this PR solving?

• Fix sampler.param shows defaultSamplingProbability instead of samplingRate #2171

Short description of the changes

• Use service's sampling param event though it has no per-operation strategies.

[codecov]

Codecov Report

Merging #2230 into master will decrease coverage by 0.01%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #2230      +/-   ##
==========================================
- Coverage   96.18%   96.16%   -0.02%     
==========================================
  Files         219      219              
  Lines       10635    10640       +5     
==========================================
+ Hits        10229    10232       +3     
- Misses        351      352       +1     
- Partials       55       56       +1     

Impacted Files	Coverage Δ
...in/sampling/strategystore/static/strategy_store.go	100.00% <100.00%> (ø)
cmd/query/app/server.go	91.78% <0.00%> (-2.74%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 0378e5b...c3e6f1a. Read the comment docs.

[yurishkuro]

is it possible that newStore.serviceStrategies[s.Service].ProbabilisticSampling == nil here?

[yurishkuro]

For example, in plugin/sampling/strategystore/static/fixtures/operation_strategies.json there is

    {
      "service": "bar",
      "type": "ratelimiting",
      "param": 5,
      ...

[yurishkuro]

Side node: this type was a mistake introduced in the past:

// operationStrategy defines an operation specific sampling strategy.
type operationStrategy struct {
        Operation string `json:"operation"`
        strategy
}

because operations can only use probabilistic strategy (example), so this generalized schema allows things that clients can't understand. Similarly, if service-level strategy contains per-operation instructions, then the top-level strategy can only be probabilistic (default sampler in the example).