Distributed query plan cache keys include the Router version number

[SimonSapin]

If you have enabled Distributed query plan caching, starting with this release and going forward you should anticipate additional cache regeneration cost when updating between any Router versions.

More often than not, an Apollo Router release may contain changes that affect what query plans are generated or how they’re represented. To avoid using outdated entries from distributed cache, the cache key includes a counter that was manually incremented with relevant data structure or algorithm changes. Instead the cache key now includes the Router version number, so that different versions will always use separate cache entries.

---

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. ↩

[svc-apollo-docs]

✅ Docs Preview Ready

No new or changed pages found.

[router-perf]

CI performance tests

• connectors-const - Connectors stress test that runs with a constant number of users
• const - Basic stress test that runs with a constant number of users
• demand-control-instrumented - A copy of the step test, but with demand control monitoring and metrics enabled
• demand-control-uninstrumented - A copy of the step test, but with demand control monitoring enabled
• enhanced-signature - Enhanced signature enabled
• events - Stress test for events with a lot of users and deduplication ENABLED
• 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_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_callback - Stress test for events with a lot of users and deduplication ENABLED in callback mode
• events_without_dedup - Stress test for events with a lot of users and deduplication DISABLED
• events_without_dedup_callback - Stress test for events with a lot of users and deduplication DISABLED using callback mode
• extended-reference-mode - Extended reference mode enabled
• large-request - Stress test with a 1 MB request payload
• no-tracing - Basic stress test, no tracing
• reload - Reload test over a long period of time at a constant rate of users
• step-jemalloc-tuning - Clone of the basic stress test for jemalloc tuning
• step-local-metrics - Field stats that are generated from the router rather than FTV1
• step-with-prometheus - A copy of the step test with the Prometheus metrics exporter enabled
• step - Basic stress test that steps up the number of users over time
• xlarge-request - Stress test with 10 MB request payload
• xxlarge-request - Stress test with 100 MB request payload

[SimonSapin]

I put this (though filename prefix) in the "Configuration" section of the changelog, but none of the options in xtask changeset create seemed like a great fit.

CHANGELOG.md has some > [!IMPORTANT] sections, is that done manually when editing for a release?

[lrlna]

we do that manually in the release PR, since it has to go in the top of the changelog

[SimonSapin]

Should this discuss the possibility a staged rollout to start warming up new cache entries while the previous version still serves most traffic? I didn’t want to sound like "draw the rest of the owl"

[lrlna]

I think "drawing the rest of the owl" can be added to distributed query plan caching docs themselves. It will be a nice addition, and people can search the docs with a bit more ease than a changelog.

[SimonSapin]

I think this doesn’t need to mention CACHE_KEY_VERSION v.s. FEDERATION_VERSION so I tried to pick a relevant level of detail.

[lrlna]

I would put this paragraph first. The point is that almost every release has had a cache key change, which comes with the additional cache regeneration cost. If people read the first paragraph, they'll think this is more than what's currently happening, but it effectively isn't.

[lrlna]

we do that manually in the release PR, since it has to go in the top of the changelog

[lrlna]

I think "drawing the rest of the owl" can be added to distributed query plan caching docs themselves. It will be a nice addition, and people can search the docs with a bit more ease than a changelog.

[lrlna]

I would put this paragraph first. The point is that almost every release has had a cache key change, which comes with the additional cache regeneration cost. If people read the first paragraph, they'll think this is more than what's currently happening, but it effectively isn't.

[lrlna]

appreciate the format! with env var so we don't have to update these tests with every release 🙏