[DO-NOT-MERGE-YET] feat(complexity): add session routing observability

[Madhuvod]

Summary

Adds session observability to complexity-routed requests. When a governance routing rule evaluates a complexity tier through session policy (pinned or cache-aware), four new fields are now recorded on the request log: a tenant-scoped session hash, the session mode, the tier source (classified, memoised, or held), and the cumulative tier-switch count. These fields are surfaced in the log detail view and are filterable from the sidebar and API.

Changes

• Added four new BifrostContextKey constants (ComplexitySessionID, ComplexitySessionMode, ComplexitySessionTierSource, ComplexitySessionSwitchCount) to the schema.
• Introduced ComplexitySessionLogFields as an embedded struct on logstore.Log, keeping database columns and JSON response fields flat.
• Added a migrationAddComplexitySessionColumns migration that idempotently adds the four nullable columns to the logs table, with partial indexes on the two low-cardinality filter columns (complexity_session_mode, complexity_session_tier_source). The session ID column is intentionally unindexed as it is display/export-only.
• Extended SearchFilters with ComplexitySessionModes and ComplexitySessionTierSources, wired through applyFilters, canUseMatViewFilters (both new filters force the raw query path), and listSelectColumns.
• Replaced duplicated inline complexity-field assignment in the logging plugin with a single applyComplexityContextToEntry helper that covers both the normal completion path and the minimal-error path, preventing future drift.
• Added publishComplexitySessionLogContext and publishSessionClassificationTelemetry in the governance plugin to emit session context into the request context from within the lazy complexity closure only—requests whose rules never inspect complexity_tier receive no session log fields.
• persistInitialSessionTier now returns the switch count and a known/unknown flag so callers can propagate it to telemetry without an extra store read.
• Extracted parseComplexityFilters in the HTTP transport to deduplicate filter parsing across the log search, stats, and histogram endpoints, and added complexity_session_modes and complexity_session_tier_sources query parameters.
• Added ComplexitySessionTierSource type, COMPLEXITY_SESSION_LOG_MODE_VALUES, COMPLEXITY_SESSION_TIER_SOURCE_VALUES, and COMPLEXITY_SESSION_TIER_SOURCE_LABELS to the UI type system.
• Added a ComplexitySessionFilter sidebar section with mode and tier-source checkboxes, wired into URL state and the API filter builder.
• Extended the log detail view to display session ID (with copy button), session mode, tier source, and switch count when present.

Type of change

• Bug fix
• Feature
• Refactor
• Documentation
• Chore/CI

Affected areas

• Core (Go)
• Transports (HTTP)
• Providers/Integrations
• Plugins
• UI (React)
• Docs

How to test

# Core/Transports
go test ./framework/logstore/... ./plugins/governance/... ./plugins/logging/... ./transports/bifrost-http/handlers/...

# UI
cd ui
pnpm i
pnpm build

Enable a routing rule that references complexity_tier with session policy (pinned or cache_aware) configured. Send several requests in the same session. Verify that:

• The log detail view shows the session ID, mode, tier source, and switch count.
• Filtering by complexity_session_modes=pinned or complexity_session_tier_sources=held returns only matching rows.
• Requests whose rules never evaluate complexity_tier have no session fields on their log entries.
• Re-running the migration against an already-migrated database completes without error and leaves existing rows intact.

Breaking changes

• Yes
• No

Related issues

Security considerations

The session ID stored in the log is a hash of the already tenant-namespaced store key, not the raw caller identity, harness value, or fingerprint. The separately hashed tenant component in the key is not exposed. No PII reaches the log column.

Checklist

• I read docs/contributing/README.md and followed the guidelines
• I added/updated tests where appropriate
• I updated documentation where needed
• I verified builds succeed (Go and UI)
• I verified the CI pipeline passes locally if applicable

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 23f0957e-a7d8-4b1b-a4cb-80d4d12bd2e6

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[Madhuvod]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• [DO-NOT-MERGE-YET] add llm classifier fallback #6074
• [DO-NOT-MERGE-YET] fix: use correct "latest" route when fallbacks are configured #6061
• [DO-NOT-MERGE-YET] feat(complexity): add session routing observability #6045 👈 (View in Graphite)
• [DO-NOT-MERGE-YET] complexity cache-aware session routing integration #6043
• [DO-NOT-MERGE-YET] complexity cache-aware session tier policy #6042
• [DO-NOT-MERGE-YET] complexity session route observations #6029
• [DO-NOT-MERGE-YET] complexity session routing and store readiness #6028
• [DO-NOT-MERGE-YET] sessions ui #6018
• [DO-NOT-MERGE-YET] feat(complexity): add session state foundation and identity resolution #6015
• [DO-NOT-MERGE-YET] feat(complexity): add session state contract and identity resolution #6009
• [DO-NOT-MERGE-YET] session routing config and sticky targets #5901
• semantic router: incremental embedding + embed usage accounting #5896 : 1 other dependent PR (#5924 )
• semantic router ui revamp #5865
• move complexity exemplars to configstore with backfill migration #5842
• feat(complexity): harness-aware complexity input extraction for cc and codex #5810
• server: export ObserveWarmupRoutingEmbedding for enterprise #5791
• weaviate + redis adapter fixes #5790
• semantic routing ui #5743
• semantic router: dimension probe + message history window #5748
• semantic cache: min similarity + fallback logging #5742
• smenatic router: budgeting+telemetry management #5686
• semantic routing #5: added semantic classifier for complexity router + handlers changes #5680
• semantic router #3: embedding executor #5656
• semantic router #2: semantic routing config #5655
• semantic router #1: add chromem backend #5628
• complexity : reducing keyword lists from 4 -> 3 lists #5602
• helm and docs changes #5530
• update complexity boundaries #5529
• expose tier/mechanism on every surface #5499
• structured logging #5497
• complexity tier merge #5494
• dev

This stack of pull requests is managed by Graphite. Learn more about stacking.