fix(postgres): treat missing incremental field as non-retryable

[Gilbert09]

Problem

Error tracking surfaced a live ValueError: incremental_field and incremental_field_type can't be None from the Postgres data-warehouse import source (issue). It fires in _build_query / _build_count_query in postgres.py before any SQL is emitted.

The trigger is a schema whose sync type wants an incremental cursor (incremental or append) but has no incremental_field stored in its config. import_data_sync passes should_use_incremental_field=True with incremental_field=None, and the query builders raise.

That stored config is fixed until the customer changes it, so the failure is deterministic. Today it surfaces as a bare ValueError that stays retryable, so every attempt re-hits the same wall, burns the retry budget, and files repeated error-tracking noise, while the customer gets no guidance.

Changes

Classify the error as non-retryable in PostgresSource.get_non_retryable_errors and map it to an actionable message: the table is set to sync incrementally but has no incremental field configured, so choose one or switch to full table replication, then re-enable the sync.

This mirrors the existing No primary key defined for table entry — same class of user-config inconsistency. I checked the reset-pipeline path: it clears the watermark but not incremental_field, so this is a persistent config state rather than a transient reset artifact, which is why non-retryable is the right call rather than leaving it to retry.

No change to the query builders or the raise sites themselves — the message they raise is already stable, and the friendly text is surfaced through the classifier the same way every other entry is.

How did you test this code?

Added test_missing_incremental_field_is_non_retryable, parameterized over both _build_query and _build_count_query. It drives the real raise sites (rather than hardcoding the string) so a future message change that breaks the classifier key is caught, then asserts the error is classified non-retryable and surfaces an actionable message.

I (Claude) could not run the suite in this environment — Python deps aren't installed here (no pytest/django). I ran ruff check and ruff format --check (both pass) and py_compile on both edited files. The new test should run under CI.

🤖 Agent context

Autonomy: Fully autonomous

I (Claude) triaged the error-tracking issue via the PostHog MCP tools, traced the call path from import_data_sync → PostgresSource.source_for_pipeline → postgres_source → _build_query, and confirmed the failure originates in this source's code. Invoked /writing-tests before adding the test.

Decision: treated this as a user/upstream config error (non-retryable) rather than a code bug, because there is nothing sensible the code can do without an incremental field and retrying can never succeed. Considered fixing at the raise site but the classifier is the established pattern for surfacing actionable messages here, so I kept the change scoped to get_non_retryable_errors. Confirmed no open PR already covers this — the closest (#68293) handles incremental field type drift and explicitly skips the None case.

---

Created with PostHog Code

[github-actions]

Hey @Gilbert09! 👋

It looks like your git author email on this PR isn't your @posthog.com address (owerstom@gmail.com). Since you're on the PostHog team, it's worth pointing your local git author email at your @posthog.com address. Why it matters:

• Consistent work identity in git history — internal tooling that attributes commits to team members keys off your @posthog.com address.
• Keeps team contributions easy to tell apart from external community ones when scanning history.

You can fix it for this repo with:

git config user.email "you@posthog.com"

Or set it globally with git config --global user.email "you@posthog.com". No need to redo this PR — just a nudge for next time. 🙂

[stamphog]

Small, additive fix to an error classifier mapping a known ValueError to non-retryable with an actionable message; verified the exact error string matches the real raise sites. No risky territory touched, author is on the owning team, and tests cover the new behavior.

• Author wrote 0% of the modified lines and has 47 merged PRs in these paths (familiarity MODERATE).
• 👍 on the PR from hex-security-app[bot].

Gate mechanics and policy version

Gate		Result
prerequisites	✓	all clear
deny-list	✓	no deny categories matched
size	✓	11L, 1F substantive, 34L/2F incl. docs/generated/snapshots — within ceiling
tier	✓	T1-agent / T1b-small (34L, 2F, single-area, fix)
stamphog 2.0.0b3		.stamphog/policy.yml @ a42508e · reviewed head 0759436

[trunk-io]

     

View Full Report ↗︎ ⋅ Docs

[deployment-status-posthog]

Deploy status

Environment	Status	Deployed At	Workflow
dev	✅ Deployed	2026-07-14 18:35 UTC	Run
prod-us	✅ Deployed	2026-07-14 18:52 UTC	Run
prod-eu	✅ Deployed	2026-07-14 18:50 UTC	Run