Add support for distributed tracing for DelayedJob

[sl0thentr0py]

• works for PerformableMethod
• activejob will be done more generally

part of #1904

[codecov]

Codecov Report

Merging #2233 (8485799) into master (a9fcc38) will increase coverage by 0.03%.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2233      +/-   ##
==========================================
+ Coverage   97.39%   97.43%   +0.03%     
==========================================
  Files         102      102              
  Lines        3804     3823      +19     
==========================================
+ Hits         3705     3725      +20     
+ Misses         99       98       -1     

Components	Coverage Δ
sentry-ruby	98.13% <ø> (+0.03%)	⬆️
sentry-rails	95.05% <ø> (ø)
sentry-sidekiq	94.70% <ø> (ø)
sentry-resque	92.06% <ø> (ø)
sentry-delayed_job	95.60% <100.00%> (+1.15%)	⬆️
sentry-opentelemetry	100.00% <ø> (ø)

Files	Coverage Δ
...entry-delayed_job/lib/sentry/delayed_job/plugin.rb	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

[st0012]

Instead of injecting trace data to args, can we maybe use a custom ivar? Like @_sentry_trace_data. That way the extraction will be easier and we don't need to risk breaking the call completely if the args deletion somehow goes wrong.

[sl0thentr0py]

Instead of injecting trace data to args, can we maybe use a custom ivar?

we'll need to patch delayed_job's YAML stuff here then, since ivar's are not picked up by default and won't survive the database round trip.
https://github.com/collectiveidea/delayed_job/blob/master/lib/delayed/psych_ext.rb#L4-L10

so I opted for this, which of the two do you prefer?

[st0012]

we'll need to patch delayed_job's YAML stuff here then

Ah thanks for this context. Let's first go with the current implementation then.

[st0012]

Not related: do you still use pry?

[sl0thentr0py]

nope! I've moved to debugger if available