Description
Component(s)
processor/tailsampling
Is your feature request related to a problem? Please describe.
The tail-based sampler currently evaluates all policies before making a decision. This makes it quite complicate to create a more complex policy set. E.g. set a default samplerate for all that do not match the previous policies.
I want to avoid the following pattern:
Define a rule:
# filter by route
name: route-name-policy,
type: string_attribute,
string_attribute:
{
key: http.route,
values: [/v1/name/.+],
enabled_regex_matching: true,
},
Define the rule once more, but invert the match to get all other:
{
# invert match - apply to all routes except the ones specified
name: route-name-policy,
type: string_attribute,
string_attribute:
{
key: http.route,
values: [/v1/name/.+],
enabled_regex_matching: true,
invert_match: true,
},
},
I want to add a flag that allows you to specify a set of policies that will determine whether a trace should be sampled or not. The policies are evaluated in order, and the first policy that matches will be used to determine the sample rate.
Describe the solution you'd like
I imagine a new flag, like first_policy_match, for the processor that changes the behaviour:
processors:
tail_sampling:
first_policy_match: true
decision_cache:
sampled_cache_size: 100_000
non_sampled_cache_size: 100_000
policies:
....
Describe alternatives you've considered
Using the current mechanism, but copying policies to add the invert_match flag. Which can be prone to errors and makes it hard to read.
Additional context
No response