chore: [PLA4-44976] remove routine Agentex polling logs

[samuelshih]

Linear

• https://linear.app/scale-epd/issue/PLA4-44976

Finding

Agentex emits two high-frequency INFO statements on normal hot paths:

• Reading messages from Redis stream <topic>, last_id: <id> before every blocking Redis read.
• Checking status of agent <id> via <url> before every health probe.

This change removes only those pre-operation success-path messages. Redis exceptions, parse warnings, non-200 health responses, unhealthy responses, agent-ID mismatches, non-JSON responses, request failures, and status changes remain logged.

Datadog evidence

Frozen window: 2026-09-04 20:08:12 UTC through 2026-09-11 20:08:12 UTC.

• Combined matching logs and seven-day volume: 69,824,740 events
• Redis polling logs: approximately 59.70M events
• Successful health-check logs: 10,125,911 INFO events
• All Agentex indexed-event volume: 160.38M events
• All Agentex ingested-byte volume: 277.54 GB

The Redis message text explicitly says INFO, but Datadog classified 370,813 of the combined matches as status:error. This removes that misleading error noise while preserving actual error statements.

Estimated savings

Contract rates: $0.98 per million indexed events (7-day retention) and $0.10 per ingested GB. Monthly estimates use 30.4375 / 7.

Cost	Seven-day removable usage	Monthly value
Indexed events	69,824,740	$297.54
Ingestion	120.836 GB estimated	$52.54
Total		$350.08/month ($4,200.99/year)

The ingestion estimate allocates measured Agentex bytes using the exact selector's share of service events. The full census projects 44.46B indexed events against a 26B monthly commitment and 158,475 GB ingestion against a 105,000 GB commitment, so gross usage value and projected invoice impact are currently the same.

Safety and provenance

• Removal confidence: 95/100.
• The Redis poll message dates to the public repository's initial commit on October 17, 2025.
• The health-check message entered public history in PR #39, merged November 4, 2025.
• pup monitors search returned 39 Agentex-related monitors. An exact scan of their names, queries, and messages found zero references to either removed string. Existing log monitors target 401s, 5xxs, registration failures, rate limits, and upstream failures, all preserved here.
• Security Monitoring Rules and saved Log Explorer views could not be audited because the current app key receives 403 for those APIs. The PR remains draft for owner review.

CODEOWNERS: @scaleapi/agentex-team for the entire repository.

Tests

• uv run pytest tests/unit/temporal/test_healthcheck_activities.py -q — 6 passed.
• uv run ruff check on both changed files.
• uv run ruff format --check on both changed files.
• git diff --check.

Rollout

Draft only. No deployment or reviewers assigned. After deploy, verify both exact Datadog queries fall while Redis exceptions and unhealthy/failed health probes remain present.