Experimental support for Apollo tracing over OTLP

[timbotnik]

https://apollographql.atlassian.net/browse/ROUTER-348

As the ecosystem around OpenTelemetry (OTel) has been expanding rapidly, we are evaluating a migration of Apollo's internal tracing system to use an OTel-based protocol.

In the short-term, benefits include:

• A comprehensive way to visualize the Router execution path in Studio.
• Additional spans that were previously not included in Studio traces, such as query parsing, planning, execution, and more.
• Additional metadata such as subgraph fetch details, Router idle / busy timing, and more.

Long-term, we see this as a strategic enhancement to consolidate these 2 disparate tracing systems.
This will pave the way for future enhancements to more easily plug into the Studio trace visualizer.

Configuration

This change adds a new configuration option experimental_otlp_tracing_sampler. This can be used to send
a percentage of traces via OTLP instead of the native Apollo Usage Reporting protocol. Supported values:

• always_off (default): send all traces via Apollo Usage Reporting protocol.
• always_on: send all traces via OTLP.
• 0.0 - 1.0: the ratio of traces to send via OTLP (0.5 = 50 / 50).

Note that this sampler is only applied after the common tracing sampler, for example:

Sample 1% of traces, send all traces via OTLP:

telemetry:
  apollo:
    # Send all traces via OTLP
    experimental_otlp_tracing_sampler: always_on

  exporters:
    tracing:
      common:
        # Sample traces at 1% of all traffic
        sampler: 0.01

---

Checklist

Complete the checklist (and note appropriate exceptions) before the PR is marked ready-for-review.

• Changes are compatible¹
• Documentation² completed
• Performance impact assessed and acceptable
• Tests added and passing³
  • Unit Tests
  • Integration Tests
  • Manual Tests

Exceptions

Note any exceptions here

Notes

Footnotes

1. It may be appropriate to bring upcoming changes to the attention of other (impacted) groups. Please endeavour to do this before seeking PR approval. The mechanism for doing this will vary considerably, so use your judgement as to how and when to do this. ↩

2. Configuration is an important part of many changes. Where applicable please try to document configuration examples. ↩

3. Tick whichever testing boxes are applicable. If you are adding Manual Tests, please document the manual testing (extensively) in the Exceptions. ↩

[router-perf]

CI performance tests

• reload - Reload test over a long period of time at a constant rate of users
• events_big_cap_high_rate_callback - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity using callback mode
• events_without_dedup_callback - Stress test for events with a lot of users and deduplication DISABLED using callback mode
• large-request - Stress test with a 1 MB request payload
• const - Basic stress test that runs with a constant number of users
• no-graphos - Basic stress test, no GraphOS.
• step-jemalloc-tuning - Clone of the basic stress test for jemalloc tuning
• events - Stress test for events with a lot of users and deduplication ENABLED
• events_callback - Stress test for events with a lot of users and deduplication ENABLED in callback mode
• events_big_cap_high_rate - Stress test for events with a lot of users, deduplication enabled and high rate event with a big queue capacity
• events_without_dedup - Stress test for events with a lot of users and deduplication DISABLED
• xxlarge-request - Stress test with 100 MB request payload
• xlarge-request - Stress test with 10 MB request payload
• step - Basic stress test that steps up the number of users over time

[garypen]

It looks generally good, but I'm concerned about the lock around the export future. Maybe worth taking another look at that or at least convincing me that it works as you'd expect.

[BrynCooke]

This is initialised in a couple of places now. Let's have a fn router_id() that contains ROUTER_ID.get_or_init(||Uuid::new_v4().to_string())

[timbotnik]

SGTM 👍

[timbotnik]

7d4a2ae

[BrynCooke]

If the integration test runner was used rather than the test harness then all the tests could be isolated.

[timbotnik]

Actually I haven't seen any flakiness TBH - I'll remove this line.

[timbotnik]

4ad4315