Added span and trace id to hotrod logs

[joe-elliott]

Which problem is this PR solving?

• Fixes Add TraceID to Logs in HotRod Example Application #2310

Short description of the changes

• Adds trace ID and span ID hotrod logs as requested.

Example:

2020-08-11T22:20:15.010-0400	INFO	route/client.go:54	Finding route	{"service": "frontend", "component": "route_client", "trace_id": "6b1ab6a4c91f3780", "span_id": "6b1ab6a4c91f3780", "pickup": "193,499", "dropoff": "577,322"}
2020-08-11T22:20:15.010-0400	INFO	route/client.go:54	Finding route	{"service": "frontend", "component": "route_client", "trace_id": "6b1ab6a4c91f3780", "span_id": "6b1ab6a4c91f3780", "pickup": "476,860", "dropoff": "577,322"}
2020-08-11T22:20:15.010-0400	INFO	route/client.go:54	Finding route	{"service": "frontend", "component": "route_client", "trace_id": "6b1ab6a4c91f3780", "span_id": "6b1ab6a4c91f3780", "pickup": "324,354", "dropoff": "577,322"}
2020-08-11T22:20:15.010-0400	INFO	route/server.go:71	HTTP request received	{"service": "route", "trace_id": "6b1ab6a4c91f3780", "span_id": "4a17e4c6aeb157ad", "method": "GET", "url": "/route?dropoff=577%2C322&pickup=193%2C499"}
2020-08-11T22:20:15.011-0400	INFO	route/server.go:71	HTTP request received	{"service": "route", "trace_id": "6b1ab6a4c91f3780", "span_id": "3b22d1cc6f90e8c", "method": "GET", "url": "/route?dropoff=577%2C322&pickup=476%2C860"}
2020-08-11T22:20:15.011-0400	INFO	route/server.go:71	HTTP request received	{"service": "route", "trace_id": "6b1ab6a4c91f3780", "span_id": "12de05f88735056f", "method": "GET", "url": "/route?dropoff=577%2C322&pickup=324%2C354"}

I know this was assigned to @ghostsquad, but it's been over a month so I took a stab at it.

[ghostsquad]

oooh, my bad, ya I've been busy with other things! sorry about that.

[codecov]

Codecov Report

Merging #2384 into master will increase coverage by 0.00%.
The diff coverage is 78.26%.

@@           Coverage Diff           @@
##           master    #2384   +/-   ##
=======================================
  Coverage   95.61%   95.61%           
=======================================
  Files         205      206    +1     
  Lines       10529    10547   +18     
=======================================
+ Hits        10067    10085   +18     
+ Misses        395      394    -1     
- Partials       67       68    +1     

Impacted Files	Coverage Δ
...el/converter/thrift/jaeger/sampling_from_domain.go	100.00% <ø> (ø)
cmd/agent/app/builder.go	96.00% <25.00%> (-4.00%)	⬇️
cmd/agent/app/servers/thriftudp/socket_buffer.go	60.00% <60.00%> (ø)
cmd/agent/app/flags.go	93.33% <100.00%> (+0.47%)	⬆️
cmd/agent/app/servers/tbuffered_server.go	100.00% <100.00%> (ø)
cmd/agent/app/servers/thriftudp/transport.go	100.00% <100.00%> (ø)
pkg/config/tlscfg/flags.go	100.00% <100.00%> (ø)
plugin/storage/kafka/options.go	94.05% <100.00%> (+0.37%)	⬆️
...lugin/sampling/strategystore/adaptive/processor.go	100.00% <0.00%> (+0.79%)	⬆️
... and 2 more

---

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 6e04d9d...6624f59. Read the comment docs.

[yurishkuro]

lgtm.

Internally at Uber we have an extension to zap.Logger that supports the notion of context extractors, e.g.

type ContextFieldExtractor = func(context.Context) zapcore.Field

and is configured with the extractor defined in Jaeger client "github.com/uber/jaeger-client-go/log/zap".Trace 

Then instead of For(ctx) and a custom logger as we do here users can use normal zap.Logger with a field Context(ctx), defined as below:

const (
   // n.b. This is used in conjunction with the skip field produced by Context() to indicate that
   //      this is a special-cased field, enabling zapfx to detect context fields and subsequently
   //      run any configured ContextFieldExtractors. When used, the corresponding zapcore.Field
   //      looks similar to:
   //
   //          zapcore.Field{
   //            Type:      zapcore.SkipType,
   //            Integer:   _contextFieldIndicator,
   //            Interface: ctx,
   //          }
   _contextFieldIndicator int64 = 1<<8 - 1
 )
    
   // Context wraps ctx as a zapcore.Field, suitable for use with calls to zap.Logger. It may only be
   // used by loggers provided by zapfx; other loggers will panic due to its internal field type.
   //
   // Importantly, Context differs from other zapcore.Field types in that it does not guarantee a 1:1
   // ratio of provided:encoded fields. Specifically, it:
   //
   //   - May not result in an encoded field, in cases in which there are no configured underlying
   //     context extractors or where context baggage does not contain the necessary context keys;
   //   - May result in more than one encoded field, in cases in which multiple context extractors
   //     find keys and produce fields; and
   //   - Does not indicate a concrete field key or value at log time, as key:value pairs of any
   //     resulting field(s) are not known until the configured extractors have run just prior to log
   //     entry encoding.
 func Context(ctx context.Context) zapcore.Field {
     return zapcore.Field{
       Type:      zapcore.SkipType,
       Integer:   _contextFieldIndicator,
       Interface: ctx,
     }
 }

Unfortunately, it never made it to OSS.