Pydantic `pydantic` weekly project plan (Sydney Runkle) -

The Fix

Improves the import times for modules in Pydantic by using caching, specifically for the `BaseModel` import.

Based on closed pydantic/pydantic issue #10055 · PR/commit linked

Production note: Watch p95/p99 latency and retry volume; timeouts can turn into retry storms and duplicate side-effects.

Open PR/Commit

@@ -2050,10 +2050,15 @@ def handle_ref_overrides(self, json_schema: JsonSchemaValue) -> JsonSchemaValue:
 
     def get_schema_from_definitions(self, json_ref: JsonRef) -> JsonSchemaValue | None:
-        def_ref = self.json_to_defs_refs[json_ref]
-        if def_ref in self._core_defs_invalid_for_json_schema:
-            raise self._core_defs_invalid_for_json_schema[def_ref]

fix.md

Option A — Apply the official fix\nImproves the import times for modules in Pydantic by using caching, specifically for the `BaseModel` import.\nWhen NOT to use: This fix should not be used if the application does not require external JSON Schema references.\n\n

Why This Fix Works in Production

Trigger: Thanks to @changhc for the sustained work on complex number support!
Mechanism: Pydantic throws a KeyError when WithJsonSchema is used with a $ref that defines an external JSON Schema via an https link

Production impact:

If left unfixed, tail latency can spike under load and surface as timeouts/retries (amplifying incident impact).

Why This Breaks in Prod

Pydantic throws a KeyError when WithJsonSchema is used with a $ref that defines an external JSON Schema via an https link
Production symptom (often without a traceback): Thanks to @changhc for the sustained work on complex number support!

Proof / Evidence

GitHub issue: #10055
Fix PR: https://github.com/pydantic/pydantic/pull/9863
Reproduced locally: No (not executed)
Last verified: 2026-02-12
Confidence: 0.60
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.71

Discussion

High-signal excerpts from the issue thread (symptoms, repros, edge-cases).

Jump to Sources Open on GitHub

“Hi Syndey, thank you for posting weekly reports”

@changhc · 2024-08-06 · source

“@changhc, Sure thing! In the meantime, you can filter by help wanted or good first issue. That being said, I'll go through the issues this…”

@sydney-runkle · 2024-08-06 · source

“Managed to make some forward progress on the union fixes with the following PRs: * https://github.com/pydantic/pydantic-core/pull/1398 * https://github.com/pydantic/pydantic-core/pull/1397 Would still like to implement some sort…”

@sydney-runkle · 2024-08-09 · source

Failure Signature (Search String)

Thanks to @changhc for the sustained work on complex number support!
Nothing at the moment, but I'm sure this will change. Ping me if you're looking for a review!

Copy-friendly signature

signature.txt

Failure Signature
-----------------
Thanks to @changhc for the sustained work on complex number support!
Nothing at the moment, but I'm sure this will change. Ping me if you're looking for a review!

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
Thanks to @changhc for the sustained work on complex number support!
Nothing at the moment, but I'm sure this will change. Ping me if you're looking for a review!

What Broke

Users experience KeyErrors when using external JSON Schema references, causing disruptions in functionality.

Why It Broke

Pydantic throws a KeyError when WithJsonSchema is used with a $ref that defines an external JSON Schema via an https link

Fix Options (Details)

Option A — Apply the official fix

Improves the import times for modules in Pydantic by using caching, specifically for the `BaseModel` import.

When NOT to use: This fix should not be used if the application does not require external JSON Schema references.

Option D — Guard side-effects with OnceOnly Guardrail for side-effects

Mitigate duplicate external side-effects under retries/timeouts/agent loops by gating the operation before calling external systems.

Place OnceOnly between your code/agent and real side-effects (Stripe, emails, CRM, APIs).
Use a stable key per side-effect (e.g., customer_id + action + idempotency_key).
Fail-safe: configure fail-open vs fail-closed based on blast radius and spend risk.

Show example snippet (optional)

onceonly.py

from onceonly import OnceOnly
import os

once = OnceOnly(api_key=os.environ["ONCEONLY_API_KEY"], fail_open=True)

# Stable idempotency key per real side-effect.
# Use a request id / job id / webhook delivery id / Stripe event id, etc.
event_id = "evt_..."  # replace
key = f"stripe:webhook:{event_id}"

res = once.check_lock(key=key, ttl=3600)
if res.duplicate:
    return {"status": "already_processed"}

# Safe to execute the side-effect exactly once.
handle_event(event_id)

See OnceOnly SDK

When NOT to use: Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

Fix reference: https://github.com/pydantic/pydantic/pull/9863

Last verified: 2026-02-12. Validate in your environment.

When NOT to Use This Fix

This fix should not be used if the application does not require external JSON Schema references.
Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

Did This Fix Work in Your Case?

Quick signal helps us prioritize which fixes to verify and improve.

Prevention

Add a CI check that diffs key outputs after upgrades (OpenAPI schema snapshots, JSON payload shapes, CLI output).
Upgrade behind a canary and run integration tests against the canary before 100% rollout.

Related Issues

No related fixes found.

Cluster: pydantic:latency Pydantic hub Pydantic best practices All hubs All clusters

Related clusters: Configuration error Data consistency OpenAPI schema

Sources

We don’t republish the full GitHub discussion text. Use the links above for context.

Pydantic `pydantic` weekly project plan (Sydney Runkle) - August 5th, 2024 (Fix)