feat(seer): Shard night shift triage into per-chunk feature runs

[trevor-e]

Night shift dispatched all scored candidates to one Seer feature run, which degrades the triage agent on large sets. This shards candidates into chunks of seer.night_shift.shard_size (default 5), dispatching each chunk as its own feature run — so max_candidates=15 triggers 3 triage agents of 5.

• New SeerNightShiftRunShard model (migration 0019) makes run → SeerRun one-to-many; one shard per chunk.
• Delivery resolves the run via a shard's SeerRun uuid, falling back to the legacy seer_run FK for pre-shard runs.
• Legacy seer_run FK kept (first shard) for the transition; backfilled and dropped in follow-up PRs.

[github-actions]

This PR has a migration; here is the generated SQL for src/sentry/seer/migrations/0019_add_night_shift_run_shard.py

for 0019_add_night_shift_run_shard in seer

--
-- Create model SeerNightShiftRunShard
--
CREATE TABLE "seer_nightshiftrunshard" ("id" bigint NOT NULL PRIMARY KEY GENERATED BY DEFAULT AS IDENTITY, "date_updated" timestamp with time zone NOT NULL, "date_added" timestamp with time zone NOT NULL, "extras" jsonb DEFAULT '{}'::jsonb NOT NULL, "run_id" bigint NOT NULL, "seer_run_id" bigint NULL UNIQUE);
ALTER TABLE "seer_nightshiftrunshard" ADD CONSTRAINT "seer_nightshiftrunsh_run_id_574f3dc7_fk_seer_nigh" FOREIGN KEY ("run_id") REFERENCES "seer_nightshiftrun" ("id") DEFERRABLE INITIALLY DEFERRED NOT VALID;
ALTER TABLE "seer_nightshiftrunshard" VALIDATE CONSTRAINT "seer_nightshiftrunsh_run_id_574f3dc7_fk_seer_nigh";
ALTER TABLE "seer_nightshiftrunshard" ADD CONSTRAINT "seer_nightshiftrunshard_seer_run_id_aa8616a4_fk_seer_seerrun_id" FOREIGN KEY ("seer_run_id") REFERENCES "seer_seerrun" ("id") DEFERRABLE INITIALLY DEFERRED NOT VALID;
ALTER TABLE "seer_nightshiftrunshard" VALIDATE CONSTRAINT "seer_nightshiftrunshard_seer_run_id_aa8616a4_fk_seer_seerrun_id";
CREATE INDEX CONCURRENTLY "seer_nightshiftrunshard_run_id_574f3dc7" ON "seer_nightshiftrunshard" ("run_id");

[trevor-e]

This is to be backwards compatible until I perform a data migration.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

^{Reviewed by Cursor Bugbot for commit b5a6802. Configure here.}

[sentry]

Bug: The seer.night_shift.shard_size option lacks validation. If set to 0, the chunked function will group all items into a single shard, defeating the purpose of sharding.
_{Severity: MEDIUM}

Suggested Fix

Add validation to the seer.night_shift.shard_size option registration in src/sentry/options/defaults.py to enforce a minimum value of 1. Alternatively, add a check in src/sentry/tasks/seer/night_shift/cron.py before calling chunked to handle a shard_size of 0 or less, perhaps by falling back to the default value or raising an error.

Prompt for AI Agent

Review the code at the location below. A potential bug has been identified by an AI
agent. Verify if this is a real issue. If it is, propose a fix; if not, explain why it's
not valid.

Location: src/sentry/tasks/seer/night_shift/cron.py#L539

Potential issue: The `seer.night_shift.shard_size` option, which is modifiable by
operators, lacks validation to prevent non-positive values. If an operator sets this
value to `0`, the `chunked` utility at `src/sentry/tasks/seer/night_shift/cron.py:539`
will not create multiple small shards. Instead, it will silently create a single, large
chunk containing all items. This defeats the purpose of the sharding logic, which is to
prevent performance degradation in the triage agent by processing large candidate sets.
The result is that all candidates are sent to the agent at once.

_{Did we get this right? 👍 / 👎 to inform future reviews.}