We are in the process to try out the opentelemetry colletor with tail sampling processor.
Our goal is to sample 10% of traces for each service.

I already tried the 10% of all traces, but sometimes it takes quite a significant time that a trace for a specific service is sampled.
For example:

      policies:
        [
          {
            name: sampling-policy,
            type: probabilistic,
            probabilistic: { sampling_percentage: 10 },
          }
        ]

Service_1 had 100.000 traces last 3h
Service_2 had 10 traces in the last 3h
Result: Only traces from Service_1 are sampled.

So we need to do something like this in our opinion:

policies: [
      {
        # Rule 1: sample 10% of service_1
        name: service_1_sampling_policy,
        type: and,
        and:
          {
            and_sub_policy:
              [
                {
                  # filter by service name
                  name: service-name-policy,
                  type: string_attribute,
                  string_attribute:
                    {
                      key: service.name,
                      values: [service-1],
                    },
                },
                {
                  # apply probabilistic sampling
                  name: probabilistic-policy,
                  type: probabilistic,
                  probabilistic: { sampling_percentage: 0.1 },
                },
              ],
          },
      },
      {
        # Rule 2: sample 10% of service_2
        name: service_2_sampling_policy,
        type: and,
        and:
          {
            and_sub_policy:
              [
                {
                  # filter by service name
                  name: service-name-policy,
                  type: string_attribute,
                  string_attribute:
                    {
                      key: service.name,
                      values: [service-2],
                    },
                },
                {
                  # apply probabilistic sampling
                  name: probabilistic-policy,
                  type: probabilistic,
                  probabilistic: { sampling_percentage: 0.1 },
                },
              ],
          },
      },
]

Now the Problem is we have a lot of services and it can happen that service names are change etc. So manual maintaining this list is not possible.

Is there a possibility to do something like:

policies: [
{
  for each service in service.name:
    {
        # Rule 1: sample 10% of each service
        name: service_sampling_policy,
        type: and,
        and:
          {
            and_sub_policy:
              [
                {
                  # filter by service name
                  name: service-name-policy,
                  type: string_attribute,
                  string_attribute:
                    {
                      values: service_name,
                    },
                },
                {
                  # apply probabilistic sampling
                  name: probabilistic-policy,
                  type: probabilistic,
                  probabilistic: { sampling_percentage: 0.1 },
                },
              ],
          },
      },
}

So the services are sampled independently of each other.