BE-742: Keep the messages Postgres sends outside a statement's results by TimDiekmann · Pull Request #9636 · hashintel/hash

TimDiekmann · 2026-09-09T20:16:37Z

🌟 What is the purpose of this PR?

Postgres reports warnings and notices on the same wire as query results, but tokio_postgres delivers them on a side channel that only Connection::poll_message exposes. Awaiting the connection, which is what a pool does by default, discards every one of them. Server warnings have therefore been invisible for as long as the graph has had a pool: SET LOCAL outside a transaction, for one, is only ever reported as a warning.

The graph now owns its pool's connection manager. Each connection runs a task that reads that channel, so what the server says becomes a tracing event under hash_graph_postgres_store::server at the level of its severity, with the SQLSTATE, detail, hint and context as fields. A termination the server initiates arrives on the same channel as a fatal error response and is recorded the same way, and a connection whose recording task has stopped is not handed out again.

On the way, the store gets a transaction API of its own instead of implementing the migration runner's traits, and deadpool-postgres goes.

🔗 Related links

BE-742 (internal)
Parent: BE-708 (internal), plan capture over the same channel, which this PR deliberately leaves out
Follow-up: BE-703 (internal), pool timeouts

🚫 Blocked by

Nothing.

🔍 What does this change?

Store transaction API (first commit): PostgresStore::transaction, the builder's isolation_level/read_only/deferrable, and commit/rollback are inherent methods. The store no longer implements hash_graph_migrations::{Context, Transaction, TransactionBuilder}, and nine test and bench files stop importing those traits to open a transaction.
Migrations crate: the runner never configured a transaction, so TransactionBuilder folds into Context::transaction and IsolationLevel moves to hash-graph-postgres-store next to TransactionOptions. hash-graph-postgres-store no longer depends on hash-graph-migrations; package.json, docs/task-dependencies.json and yarn.lock are regenerated accordingly.
connection.rs (new): ManagedConnection pairs a tokio_postgres::Client with the task driving poll_message; ConnectionManager implements deadpool::managed::Manager and rejects a connection on recycle when its client is closed or its task has finished. ConnectionError replaces deadpool_postgres::PoolError as StorePool::Error.
pool.rs: PooledConnection is a newtype over deadpool's Object, which keeps deadpool a private dependency. PostgresStorePool::new never connected and is no longer async; the callers drop their .await.
deadpool-postgres is removed. Only its Config was ever used: the statement cache is opt-in and was never opted into, and its transaction wrappers exist to offer that same cache.
The SET LOCAL sites in the store already run inside transactions, so the new warning does not fire on a hot path.

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

This PR:

does not modify any publishable blocks or libraries, or modifications do not need publishing

📜 Does this require a change to the docs?

The changes in this PR:

are internal and do not require a docs change

🕸️ Does this require a change to the Turbo Graph?

The changes in this PR:

affected the execution graph, and the turbo.json's have been updated to reflect this
- hash-graph-postgres-store loses its edge to hash-graph-migrations; mise run sync:turborepo regenerated package.json and docs/task-dependencies.json, no turbo.json needed a change

⚠️ Known issues

Pool timeouts stay unbounded, as before: deadpool's defaults are all None, matching the explicit Nones the old configuration set. ConnectionError::Unavailable therefore covers a wait timeout and a missing runtime with one variant, and no pool error carries a StatusCode, so a saturated pool would answer 500 rather than 503. Unreachable until a timeout is configured; tracked in BE-703, where the variant split and the status code belong.
Server messages carry no request trace. The recording task runs outside any span on purpose, because a connection outlives the request that made the pool grow. Every event carries the connection id as a field of its own, so it survives any filter level, and that id joins it to the acquisition that created the connection, which is logged within that request's span. Attributing a message to the statement that provoked it needs a collector, which is BE-708's subject.
Session state is not reset when a connection returns to the pool, unchanged from deadpool-postgres's RecyclingMethod::Fast. The store sets session parameters only with SET LOCAL inside a transaction, and a SET LOCAL outside one now shows up as a warning, so a reset round trip per checkout is not worth its cost.

🐾 Next steps

BE-708: plan capture via auto_explain notices over this channel, with a collector per transaction.
BE-741: turning plan capture on in a deployment without a redeploy.
BE-703: pool timeouts, with the error split and the status code above.

🛡 What tests cover this?

New, in libs/@local/graph/postgres-store/tests/connection/main.rs, each installing a tracing layer that records what reaches hash_graph_postgres_store::server:
- connection_records_server_warnings: SET LOCAL outside a transaction is recorded at WARN with severity, SQLSTATE 25P01 and the connection id on the event.
- connection_records_server_notices: RAISE NOTICE … USING DETAIL, HINT is recorded at INFO with detail, hint and context, so the level mapping is pinned on more than one arm and every field the server sends is asserted.
- connection_records_notifications: LISTEN and NOTIFY on one connection record the notification at INFO with its channel.
- connection_records_server_termination: pg_terminate_backend from a second connection is recorded at ERROR with FATAL and 57P01.
- pool_replaces_terminated_connection: a pool of one connection does not hand a terminated backend out again.
- pool_reports_unreachable_database: the pool connects lazily and reports a missing database as ConnectionError::Connect with the driver's error in the report.
Existing, in tests/graph/integration/postgres/transaction.rs: options composition, defaults and savepoint nesting through the store's transaction API.
Every other integration test reaches the database through the new pool.

❓ How to test this?

Start the compose stack and run cargo nextest run -p hash-graph-postgres-store --all-features --test connection.
Run the graph with RUST_LOG=hash_graph_postgres_store::server=info and, from psql, SELECT pg_terminate_backend(pid) FROM pg_stat_activity WHERE usename = 'graph' AND state = 'idle' LIMIT 1;.
Confirm the graph logs an ERROR event with severity=FATAL and code=57P01, and that the next request succeeds on a fresh connection.

📹 Demo

No UI. The events the tests assert on are what an operator sees in the graph's log.

`Context`, `TransactionBuilder` and `Transaction` belong to the migration runner, which is generic over contexts and reaches a transaction through them. The store implemented them because it runs migrations, but then had no transaction API of its own, so every ordinary read and write in the data path went through the migration runner's contract. Opening a transaction meant importing a migrations trait, which nine test and bench files did. The setters and the entry point are now the store's own. The runner never configured a transaction, so `TransactionBuilder` loses its options and folds into `Context::transaction`, and `IsolationLevel` moves next to the options that use it. The store no longer depends on the migrations crate. Co-authored-by: Claude <noreply@anthropic.com>

`tokio_postgres` delivers warnings, notices and notifications on a side channel that only `Connection::poll_message` exposes, and awaiting the connection — which is what a pool does by default — discards every one of them. Server warnings were therefore invisible: `SET LOCAL` outside a transaction, for one, only ever says so in a warning. `ManagedConnection` drives that channel from a task of its own, so each message becomes a tracing event under `hash_graph_postgres_store::server` at the level of its severity, with the SQLSTATE, detail, hint and context as fields. A termination the server initiates arrives on the same channel as a fatal error response and is recorded the same way. A connection whose recording task has stopped counts as closed, so the pool does not hand it out again. This replaces `deadpool-postgres` with a manager over `deadpool` itself. Only `Config` was ever used from it: the statement cache is opt-in and was never opted into, and the transaction wrappers exist to offer that same cache. `PostgresStorePool::new` never connected, so it is no longer async. Co-authored-by: Claude <noreply@anthropic.com>

…rations crate

vercel · 2026-09-09T20:16:43Z

The latest updates on your projects. Learn more about Vercel for GitHub.

4 Skipped Deployments

Project	Deployment	Actions	Updated
hash	Ignored	Preview	Sep 10, 2026 11:03am UTC
hashdotdesign-tokens	Ignored	Preview	Sep 10, 2026 11:03am UTC
petrinaut	Skipped		Sep 10, 2026 11:03am UTC
petrinaut-docs	Skipped		Sep 10, 2026 11:03am UTC

codecov · 2026-09-09T20:24:10Z

Codecov Report

❌ Patch coverage is 0% with 144 lines in your changes missing coverage. Please review.
✅ Project coverage is 65.86%. Comparing base (407734a) to head (11acaae).
⚠️ Report is 12 commits behind head on main.

Files with missing lines	Patch %	Lines
...ph/postgres-store/src/store/postgres/connection.rs	0.00%	109 Missing ⚠️
...al/graph/postgres-store/src/store/postgres/pool.rs	0.00%	29 Missing ⚠️
...cal/graph/postgres-store/src/store/postgres/mod.rs	0.00%	6 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9636      +/-   ##
==========================================
- Coverage   65.89%   65.86%   -0.04%     
==========================================
  Files        1887     1888       +1     
  Lines      198443   198559     +116     
  Branches     8248     8248              
==========================================
- Hits       130773   130772       -1     
- Misses      66140    66257     +117     
  Partials     1530     1530

Flag	Coverage Δ
apps.hash-ai-worker-ts	`1.99% <ø> (ø)`
apps.hash-api	`15.35% <ø> (ø)`
apps.hash-graph	`12.62% <ø> (+0.08%)`	⬆️
blockprotocol.type-system	`38.15% <ø> (ø)`
local.claude-hooks	`0.00% <ø> (ø)`
local.harpc-client	`51.49% <ø> (ø)`
local.hash-backend-utils	`3.27% <ø> (ø)`
local.hash-graph-sdk	`10.02% <ø> (ø)`
local.hash-isomorphic-utils	`12.22% <ø> (ø)`
rust.antsi	`2.36% <ø> (ø)`
rust.error-stack	`90.81% <ø> (ø)`
rust.harpc-codec	`84.70% <ø> (ø)`
rust.harpc-net	`96.19% <ø> (-0.02%)`	⬇️
rust.harpc-tower	`67.03% <ø> (ø)`
rust.harpc-types	`0.00% <ø> (ø)`
rust.harpc-wire-protocol	`92.23% <ø> (ø)`
rust.hash-codec	`72.76% <ø> (ø)`
rust.hash-config	`81.14% <ø> (ø)`
rust.hash-graph-api	`19.71% <ø> (ø)`
rust.hash-graph-atlas	`80.36% <ø> (ø)`
rust.hash-graph-authentication	`96.02% <ø> (ø)`
rust.hash-graph-authorization	`63.14% <ø> (ø)`
rust.hash-graph-embeddings	`91.88% <ø> (ø)`
rust.hash-graph-postgres-store	`31.96% <0.00%> (-0.19%)`	⬇️
rust.hash-graph-store	`48.41% <ø> (ø)`
rust.hash-graph-temporal-versioning	`50.18% <ø> (ø)`
rust.hash-graph-types	`0.00% <ø> (ø)`
rust.hash-graph-validation	`84.71% <ø> (ø)`
rust.hash-middleware	`90.92% <ø> (ø)`
rust.hashql-ast	`89.63% <ø> (ø)`
rust.hashql-compiletest	`28.39% <ø> (ø)`
rust.hashql-core	`78.92% <ø> (ø)`
rust.hashql-diagnostics	`72.51% <ø> (ø)`
rust.hashql-eval	`79.82% <ø> (ø)`
rust.hashql-hir	`89.09% <ø> (ø)`
rust.hashql-mir	`87.92% <ø> (ø)`
rust.hashql-syntax-jexpr	`94.04% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

codspeed-hq · 2026-09-09T20:25:39Z

Merging this PR will not alter performance

⚠️

6 benchmarks measured no execution time

Nothing ran under measurement, usually because the compiler removed the code under test. These results are not comparable, so they count as unchanged.

Preventing compiler optimizations

✅ 98 untouched benchmarks

Performance Changes

	Benchmark	`BASE`	`HEAD`	Efficiency
⚠️	`as_constant`	< 1 ns	< 1 ns	N/A
⚠️	`constant_equal`	< 1 ns	< 1 ns	N/A
⚠️	`constant_not_equal`	< 1 ns	< 1 ns	N/A
⚠️	`access`	< 1 ns	< 1 ns	N/A
⚠️	`runtime_equal`	< 1 ns	< 1 ns	N/A
⚠️	`runtime_not_equal`	< 1 ns	< 1 ns	N/A

_{Comparing t/be-742-keep-the-messages-postgres-sends-outside-a-statements (11acaae) with main (4c664c0)¹}

No successful run was found on main (890b8c4) during the generation of this report, so 4c664c0 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

cursor · 2026-09-10T10:47:36Z

PR Summary

Medium Risk
Refactors the core Postgres connection pool and transaction API used by every graph CLI, server, and test; behavior changes include lazy connect, new error types, and decoupling from the migrations crate.

Overview
Replaces deadpool-postgres with a custom deadpool::managed pool so each checkout keeps a background task on tokio_postgres::Connection::poll_message. Postgres warnings, notices, notifications, and server-initiated terminations are emitted as tracing events under hash_graph_postgres_store::server (with SQLSTATE, detail, hint, context) instead of being dropped when the connection is driven only for pooling.

PostgresStorePool::new is now synchronous and connects lazily on acquire; pool failures surface as ConnectionError instead of deadpool_postgres::PoolError. PooledConnection / ManagedConnection wrap the client and recycle rejects closed backends.

hash-graph-postgres-store no longer depends on hash-graph-migrations. Transaction control moves onto PostgresStore (transaction() builder, commit / rollback, IsolationLevel in the store crate); the migrations crate’s Context API is simplified to a plain async fn transaction() without TransactionBuilder. Call sites drop .await on pool construction and stop importing migration transaction traits.

Adds tests/connection coverage for warning/notice/notification/termination logging and pool recycle behavior; updates workspace lockfiles and dependency diagrams accordingly.

^{Reviewed by Cursor Bugbot for commit 11acaae. Bugbot is set up for automated code reviews on this repo. Configure here.}

Copilot

🟡 Changes recommended

WARN-level filtering drops connection correlation, and key preserved message fields remain untested.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

This PR replaces deadpool-postgres with a custom managed PostgreSQL pool that records asynchronous server messages through tracing and exposes native store transaction APIs.

Changes:

Adds managed connections that record notices, warnings, notifications, and termination errors.
Moves transaction configuration into PostgresStore and simplifies migration traits.
Updates dependencies, callers, tests, benchmarks, and generated dependency metadata.

File summaries

File	Description
`yarn.lock`	Regenerates workspace dependency wiring.
`tests/graph/integration/postgres/transaction.rs`	Uses inherent transaction APIs.
`tests/graph/integration/postgres/lib.rs`	Adapts synchronous pool construction.
`tests/graph/benches/util.rs`	Updates benchmark pool construction.
`tests/graph/benches/representative_read/seed.rs`	Removes migration trait imports.
`tests/graph/benches/read_scaling/knowledge/linkless/entity.rs`	Removes migration trait imports.
`tests/graph/benches/read_scaling/knowledge/complete/entity.rs`	Removes migration trait imports.
`tests/graph/benches/policy/seed.rs`	Uses inherent transaction methods.
`tests/graph/benches/manual_queries/entity_queries/mod.rs`	Removes unnecessary runtime blocking.
`tests/graph/benches/graph/scenario/runner.rs`	Adapts synchronous pool construction.
`libs/@local/telemetry/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/syntax-jexpr/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/mir/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/hir/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/eval/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/compiletest/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/hashql/ast/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/graph/postgres-store/tests/principals/main.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/tests/principals/actions.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/tests/deletion/main.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/tests/connection/main.rs`	Tests server-message recording and recycling.
`libs/@local/graph/postgres-store/tests/common/mod.rs`	Extracts reusable connection configuration.
`libs/@local/graph/postgres-store/src/store/postgres/query/compile/peephole/tuple.rs`	Updates statement-cache documentation.
`libs/@local/graph/postgres-store/src/store/postgres/pool.rs`	Implements the custom managed pool wrapper.
`libs/@local/graph/postgres-store/src/store/postgres/ontology/property_type.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/store/postgres/ontology/entity_type.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/store/postgres/ontology/data_type.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/store/postgres/mod.rs`	Adds inherent transaction APIs and exports.
`libs/@local/graph/postgres-store/src/store/postgres/knowledge/entity/mod.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/store/postgres/connection.rs`	Drives and records PostgreSQL side-channel messages.
`libs/@local/graph/postgres-store/src/store/mod.rs`	Updates the store’s public exports.
`libs/@local/graph/postgres-store/src/snapshot/mod.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/permissions/mod.rs`	Removes obsolete trait import.
`libs/@local/graph/postgres-store/src/lib.rs`	Enables trait aliases.
`libs/@local/graph/postgres-store/package.json`	Removes generated migrations dependency wiring.
`libs/@local/graph/postgres-store/docs/task-dependencies.json`	Updates generated task dependencies.
`libs/@local/graph/postgres-store/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/graph/postgres-store/Cargo.toml`	Replaces `deadpool-postgres` dependencies and features.
`libs/@local/graph/migrations/src/postgres.rs`	Simplifies PostgreSQL migration transactions.
`libs/@local/graph/migrations/src/lib.rs`	Removes transaction-builder exports.
`libs/@local/graph/migrations/src/context.rs`	Folds transaction creation into `Context`.
`libs/@local/graph/migrations/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/graph/migrations-macros/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/graph/atlas/tests/route_fixture.rs`	Adapts synchronous pool construction.
`libs/@local/graph/atlas/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`libs/@local/graph/api/docs/dependency-diagram.mmd`	Regenerates dependency graph.
`Cargo.toml`	Removes the workspace `deadpool-postgres` dependency.
`Cargo.lock`	Regenerates Rust dependency resolution.
`apps/hash-graph/src/subcommand/snapshot.rs`	Adapts synchronous pool construction.
`apps/hash-graph/src/subcommand/server.rs`	Adapts synchronous pool construction.
`apps/hash-graph/src/subcommand/reindex_cache.rs`	Adapts synchronous pool construction.
`apps/hash-graph/src/subcommand/migrate.rs`	Adapts synchronous pool construction.
`apps/hash-graph/src/subcommand/atlas.rs`	Adapts synchronous pool construction.
`apps/hash-graph/src/subcommand/admin_server.rs`	Adapts synchronous pool construction.
`apps/hash-graph/docs/dependency-diagram.mmd`	Regenerates dependency graph.

Review details

Suppressed comments (1)

libs/@local/graph/postgres-store/src/store/postgres/connection.rs:263

An INFO span is disabled when the active filter is WARN (for example HASH_GRAPH_LOG_LEVEL=warn), so the WARN/ERROR server events emitted inside this instrumented future lose the span's connection field. That breaks the advertised correlation exactly at a normal production threshold. Thread id through drive/record/report, record it directly on every server event, and test this with a WARN-filtered subscriber.

            let span = tracing::info_span!(parent: None, "postgres_connection", connection = id);

Files reviewed: 53/55 changed files
Comments generated: 1
Review effort level: Balanced

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

… span An INFO span is disabled under a WARN filter, and the WARN and ERROR events recorded inside it lose the span's fields with it. The connection id is a field of each event now, so it survives whatever filter is active.

github-actions · 2026-09-10T12:22:18Z

Benchmark results

@rust/hash-graph-benches – Integrations

policy_resolution_large

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2002	$$27.8 \mathrm{ms} \pm 238 \mathrm{μs}\left({\color{gray}0.250 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$3.45 \mathrm{ms} \pm 25.6 \mathrm{μs}\left({\color{gray}-0.448 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 1002	$$12.8 \mathrm{ms} \pm 125 \mathrm{μs}\left({\color{gray}1.19 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 3314	$$43.5 \mathrm{ms} \pm 395 \mathrm{μs}\left({\color{gray}0.255 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$14.3 \mathrm{ms} \pm 131 \mathrm{μs}\left({\color{gray}1.23 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 1527	$$24.4 \mathrm{ms} \pm 238 \mathrm{μs}\left({\color{gray}1.49 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 2078	$$28.6 \mathrm{ms} \pm 232 \mathrm{μs}\left({\color{gray}-0.177 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$3.78 \mathrm{ms} \pm 29.5 \mathrm{μs}\left({\color{gray}1.44 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 1033	$$13.9 \mathrm{ms} \pm 129 \mathrm{μs}\left({\color{gray}0.925 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_medium

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 102	$$3.85 \mathrm{ms} \pm 35.3 \mathrm{μs}\left({\color{gray}-1.326 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$3.08 \mathrm{ms} \pm 24.5 \mathrm{μs}\left({\color{gray}0.400 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 52	$$3.47 \mathrm{ms} \pm 25.4 \mathrm{μs}\left({\color{gray}1.49 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 269	$$5.29 \mathrm{ms} \pm 43.8 \mathrm{μs}\left({\color{gray}-0.242 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$3.60 \mathrm{ms} \pm 22.5 \mathrm{μs}\left({\color{gray}-1.773 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 108	$$4.24 \mathrm{ms} \pm 31.3 \mathrm{μs}\left({\color{gray}-0.435 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 133	$$4.58 \mathrm{ms} \pm 36.9 \mathrm{μs}\left({\color{gray}1.17 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$3.50 \mathrm{ms} \pm 27.9 \mathrm{μs}\left({\color{gray}-0.360 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 63	$$4.25 \mathrm{ms} \pm 32.3 \mathrm{μs}\left({\color{gray}1.35 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_none

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 2	$$2.76 \mathrm{ms} \pm 18.7 \mathrm{μs}\left({\color{gray}-0.089 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.74 \mathrm{ms} \pm 18.0 \mathrm{μs}\left({\color{gray}1.25 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 2	$$2.85 \mathrm{ms} \pm 19.2 \mathrm{μs}\left({\color{gray}1.43 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 8	$$3.12 \mathrm{ms} \pm 21.3 \mathrm{μs}\left({\color{gray}1.42 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$2.93 \mathrm{ms} \pm 26.9 \mathrm{μs}\left({\color{gray}1.28 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 3	$$3.21 \mathrm{ms} \pm 24.7 \mathrm{μs}\left({\color{gray}1.02 \mathrm{\%}}\right) $$	Flame Graph

policy_resolution_small

Function	Value	Mean	Flame graphs
resolve_policies_for_actor	user: empty, selectivity: high, policies: 52	$$3.13 \mathrm{ms} \pm 26.1 \mathrm{μs}\left({\color{gray}-0.154 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: low, policies: 1	$$2.87 \mathrm{ms} \pm 23.2 \mathrm{μs}\left({\color{gray}1.77 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: empty, selectivity: medium, policies: 26	$$2.99 \mathrm{ms} \pm 17.8 \mathrm{μs}\left({\color{gray}0.094 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: high, policies: 94	$$3.59 \mathrm{ms} \pm 27.6 \mathrm{μs}\left({\color{gray}-0.066 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: low, policies: 1	$$3.15 \mathrm{ms} \pm 27.1 \mathrm{μs}\left({\color{gray}2.72 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: seeded, selectivity: medium, policies: 27	$$3.39 \mathrm{ms} \pm 25.1 \mathrm{μs}\left({\color{gray}0.343 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: high, policies: 66	$$3.57 \mathrm{ms} \pm 30.7 \mathrm{μs}\left({\color{gray}1.74 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: low, policies: 1	$$3.12 \mathrm{ms} \pm 22.1 \mathrm{μs}\left({\color{gray}1.16 \mathrm{\%}}\right) $$	Flame Graph
resolve_policies_for_actor	user: system, selectivity: medium, policies: 29	$$3.44 \mathrm{ms} \pm 23.4 \mathrm{μs}\left({\color{gray}0.298 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_complete

Function	Value	Mean	Flame graphs
entity_by_id;one_depth	1 entities	$$32.3 \mathrm{ms} \pm 206 \mathrm{μs}\left({\color{gray}1.50 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	10 entities	$$71.4 \mathrm{ms} \pm 620 \mathrm{μs}\left({\color{gray}1.59 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	25 entities	$$36.1 \mathrm{ms} \pm 250 \mathrm{μs}\left({\color{gray}3.04 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	5 entities	$$39.7 \mathrm{ms} \pm 248 \mathrm{μs}\left({\color{gray}0.814 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;one_depth	50 entities	$$41.8 \mathrm{ms} \pm 227 \mathrm{μs}\left({\color{gray}-3.873 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	1 entities	$$33.9 \mathrm{ms} \pm 193 \mathrm{μs}\left({\color{gray}0.481 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	10 entities	$$418 \mathrm{ms} \pm 1.41 \mathrm{ms}\left({\color{gray}0.404 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	25 entities	$$91.5 \mathrm{ms} \pm 709 \mathrm{μs}\left({\color{gray}-0.115 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	5 entities	$$78.9 \mathrm{ms} \pm 594 \mathrm{μs}\left({\color{gray}-0.209 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;two_depth	50 entities	$$277 \mathrm{ms} \pm 1.22 \mathrm{ms}\left({\color{gray}1.11 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	1 entities	$$10.6 \mathrm{ms} \pm 79.2 \mathrm{μs}\left({\color{gray}1.53 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	10 entities	$$10.9 \mathrm{ms} \pm 64.9 \mathrm{μs}\left({\color{gray}3.37 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	25 entities	$$10.8 \mathrm{ms} \pm 66.4 \mathrm{μs}\left({\color{gray}0.375 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	5 entities	$$10.8 \mathrm{ms} \pm 57.4 \mathrm{μs}\left({\color{gray}2.01 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id;zero_depth	50 entities	$$10.8 \mathrm{ms} \pm 90.7 \mathrm{μs}\left({\color{gray}1.78 \mathrm{\%}}\right) $$	Flame Graph

read_scaling_linkless

Function	Value	Mean	Flame graphs
entity_by_id	1 entities	$$10.6 \mathrm{ms} \pm 86.0 \mathrm{μs}\left({\color{gray}0.792 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10 entities	$$10.6 \mathrm{ms} \pm 59.6 \mathrm{μs}\left({\color{gray}0.448 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	100 entities	$$10.7 \mathrm{ms} \pm 87.0 \mathrm{μs}\left({\color{gray}2.20 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	1000 entities	$$10.8 \mathrm{ms} \pm 63.2 \mathrm{μs}\left({\color{gray}4.27 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	10000 entities	$$10.9 \mathrm{ms} \pm 71.6 \mathrm{μs}\left({\color{gray}0.143 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity

Function	Value	Mean	Flame graphs
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/block/v/1`	$$11.6 \mathrm{ms} \pm 70.3 \mathrm{μs}\left({\color{gray}4.25 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/book/v/1`	$$11.4 \mathrm{ms} \pm 73.8 \mathrm{μs}\left({\color{gray}4.37 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/building/v/1`	$$11.6 \mathrm{ms} \pm 72.9 \mathrm{μs}\left({\color{red}6.09 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/organization/v/1`	$$11.4 \mathrm{ms} \pm 75.4 \mathrm{μs}\left({\color{gray}4.49 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/page/v/2`	$$11.5 \mathrm{ms} \pm 101 \mathrm{μs}\left({\color{gray}4.64 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/person/v/1`	$$11.2 \mathrm{ms} \pm 68.3 \mathrm{μs}\left({\color{gray}2.94 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/playlist/v/1`	$$11.7 \mathrm{ms} \pm 83.0 \mathrm{μs}\left({\color{red}6.77 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/song/v/1`	$$11.3 \mathrm{ms} \pm 68.2 \mathrm{μs}\left({\color{gray}2.39 \mathrm{\%}}\right) $$	Flame Graph
entity_by_id	entity type ID: `https://blockprotocol.org/@alice/types/entity-type/uk-address/v/1`	$$11.8 \mathrm{ms} \pm 93.2 \mathrm{μs}\left({\color{red}6.81 \mathrm{\%}}\right) $$	Flame Graph

representative_read_entity_type

Function	Value	Mean	Flame graphs
get_entity_type_by_id	Account ID: `bf5a9ef5-dc3b-43cf-a291-6210c0321eba`	$$8.34 \mathrm{ms} \pm 55.0 \mathrm{μs}\left({\color{gray}2.76 \mathrm{\%}}\right) $$	Flame Graph

representative_read_multiple_entities

Function	Value	Mean	Flame graphs
entity_by_property	traversal_paths=0	0	$$57.1 \mathrm{ms} \pm 396 \mathrm{μs}\left({\color{gray}3.23 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$111 \mathrm{ms} \pm 701 \mathrm{μs}\left({\color{gray}3.21 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$63.2 \mathrm{ms} \pm 440 \mathrm{μs}\left({\color{gray}3.41 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$72.6 \mathrm{ms} \pm 527 \mathrm{μs}\left({\color{gray}2.58 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$81.8 \mathrm{ms} \pm 418 \mathrm{μs}\left({\color{gray}2.80 \mathrm{\%}}\right) $$
entity_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$89.2 \mathrm{ms} \pm 766 \mathrm{μs}\left({\color{gray}2.93 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=0	0	$$46.9 \mathrm{ms} \pm 284 \mathrm{μs}\left({\color{gray}3.59 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=255	1,resolve_depths=inherit:1;values:255;properties:255;links:127;link_dests:126;type:true	$$73.9 \mathrm{ms} \pm 503 \mathrm{μs}\left({\color{gray}0.794 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:0;link_dests:0;type:false	$$53.1 \mathrm{ms} \pm 425 \mathrm{μs}\left({\color{gray}4.08 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:0;links:1;link_dests:0;type:true	$$61.7 \mathrm{ms} \pm 475 \mathrm{μs}\left({\color{gray}3.44 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:0;properties:2;links:1;link_dests:0;type:true	$$63.1 \mathrm{ms} \pm 399 \mathrm{μs}\left({\color{gray}1.56 \mathrm{\%}}\right) $$
link_by_source_by_property	traversal_paths=2	1,resolve_depths=inherit:0;values:2;properties:2;links:1;link_dests:0;type:true	$$63.2 \mathrm{ms} \pm 585 \mathrm{μs}\left({\color{gray}2.12 \mathrm{\%}}\right) $$

scenarios

Function	Value	Mean	Flame graphs
full_test	query-limited	$$108 \mathrm{ms} \pm 847 \mathrm{μs}\left({\color{lightgreen}-5.564 \mathrm{\%}}\right) $$	Flame Graph
full_test	query-unlimited	$$120 \mathrm{ms} \pm 701 \mathrm{μs}\left({\color{lightgreen}-5.090 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-limited	$$23.2 \mathrm{ms} \pm 190 \mathrm{μs}\left({\color{red}25.5 \mathrm{\%}}\right) $$	Flame Graph
linked_queries	query-unlimited	$$519 \mathrm{ms} \pm 1.28 \mathrm{ms}\left({\color{gray}-1.042 \mathrm{\%}}\right) $$	Flame Graph

TimDiekmann and others added 3 commits September 9, 2026 18:19

BE-742: Regenerate the lockfile and task dependencies without the mig…

e62f814

…rations crate

TimDiekmann self-assigned this Sep 9, 2026

TimDiekmann temporarily deployed to pull-request September 9, 2026 20:17 — with GitHub Actions Inactive

vercel Bot deployed to Preview – petrinaut September 9, 2026 20:18 View deployment

vercel Bot deployed to Preview – petrinaut-docs September 9, 2026 20:19 View deployment

vercel Bot deployed to Preview – hash September 9, 2026 20:32 View deployment

github-advanced-security Bot found potential problems Sep 9, 2026

View reviewed changes

Comment thread libs/@local/graph/postgres-store/tests/common/mod.rs Dismissed

Comment thread tests/graph/benches/manual_queries/entity_queries/mod.rs Dismissed

Update dependency diagrams

2f247f6

vercel Bot temporarily deployed to Preview – petrinaut-docs September 10, 2026 07:59 Inactive

vercel Bot temporarily deployed to Preview – petrinaut September 10, 2026 07:59 Inactive

TimDiekmann temporarily deployed to pull-request September 10, 2026 08:00 — with GitHub Actions Inactive

TimDiekmann marked this pull request as ready for review September 10, 2026 10:47

TimDiekmann requested a review from a team as a code owner September 10, 2026 10:47

TimDiekmann requested review from a team and a balanced review from Copilot September 10, 2026 10:47

Copilot started reviewing on behalf of TimDiekmann September 10, 2026 10:48 View session

Copilot AI reviewed Sep 10, 2026

View reviewed changes

Comment thread libs/@local/graph/postgres-store/src/store/postgres/connection.rs

vercel Bot temporarily deployed to Preview – petrinaut September 10, 2026 11:02 Inactive

vercel Bot temporarily deployed to Preview – petrinaut-docs September 10, 2026 11:02 Inactive

TimDiekmann temporarily deployed to pull-request September 10, 2026 11:03 — with GitHub Actions Inactive

claude Bot mentioned this pull request Sep 10, 2026

SRE-1049: Move the Brunch chat and runbook tests out of test:unit #9643

Merged

3 tasks

Navigation Menu

Uh oh!

BE-742: Keep the messages Postgres sends outside a statement's results - #9636

BE-742: Keep the messages Postgres sends outside a statement's results#9636
TimDiekmann wants to merge 5 commits into
mainfrom
t/be-742-keep-the-messages-postgres-sends-outside-a-statements

TimDiekmann commented Sep 9, 2026 •

edited

Loading

Uh oh!

vercel Bot commented Sep 9, 2026 •

edited

Loading

Uh oh!

codecov Bot commented Sep 9, 2026 •

edited

Loading

Uh oh!

codspeed-hq Bot commented Sep 9, 2026 •

edited

Loading

Uh oh!

Uh oh!

Uh oh!

cursor Bot commented Sep 10, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

github-actions Bot commented Sep 10, 2026

@rust/hash-graph-benches – Integrations

policy_resolution_large

policy_resolution_medium

policy_resolution_none

policy_resolution_small

read_scaling_complete

read_scaling_linkless

representative_read_entity

representative_read_entity_type

representative_read_multiple_entities

scenarios

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

3 participants

Uh oh!

Conversation

TimDiekmann commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🌟 What is the purpose of this PR?

🔗 Related links

🚫 Blocked by

🔍 What does this change?

Pre-Merge Checklist 🚀

🚢 Has this modified a publishable library?

📜 Does this require a change to the docs?

🕸️ Does this require a change to the Turbo Graph?

⚠️ Known issues

🐾 Next steps

🛡 What tests cover this?

❓ How to test this?

📹 Demo

Uh oh!

vercel Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

codspeed-hq Bot commented Sep 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Performance Changes

Footnotes

Uh oh!

Uh oh!

Uh oh!

cursor Bot commented Sep 10, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Summary

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

🟡 Changes recommended

Suppressed comments (1)

Uh oh!

Uh oh!

github-actions Bot commented Sep 10, 2026

Benchmark results

@rust/hash-graph-benches – Integrations

policy_resolution_large

policy_resolution_medium

policy_resolution_none

policy_resolution_small

read_scaling_complete

read_scaling_linkless

representative_read_entity

representative_read_entity_type

representative_read_multiple_entities

scenarios

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

3 participants

TimDiekmann commented Sep 9, 2026 •

edited

Loading

vercel Bot commented Sep 9, 2026 •

edited

Loading

codecov Bot commented Sep 9, 2026 •

edited

Loading

codspeed-hq Bot commented Sep 9, 2026 •

edited

Loading

cursor Bot commented Sep 10, 2026 •

edited

Loading