Replace fixed-interval LLM log polling with exponential backoff

[Copilot]

The /llm_log poll ran on a fixed setInterval forever — accumulating stale intervals across sessions, navigation, and generation lifecycle events.

Approach

Replaced setInterval with a self-rescheduling setTimeout pattern. After each poll, the next one is scheduled based on how many consecutive "no new entries" results have been seen:

Consecutive stale polls	Next poll delay
0–1	15 s
2	30 s
3	60 s
4	2 min
5	5 min
6+	10 min

Any poll that finds new entries resets the counter immediately back to 15 s.

Key changes (static/js/script.js)

• _logPollTimeout / _logPollSameCount / _LOG_POLL_DELAYS — replace _logPollInterval; track backoff state
• _scheduleLogPoll() — cancels any pending timeout, picks the delay for current _logPollSameCount, fires next setTimeout
• pollLLMLog() — tracks foundNew per poll; resets or increments _logPollSameCount before calling _scheduleLogPoll(); error path also increments backoff
• Clear-log handler — resets _logPollSameCount = 0 so polling snaps back to 15 s after a manual clear

var _LOG_POLL_DELAYS = [15000, 15000, 30000, 60000, 120000, 300000, 600000];

function _scheduleLogPoll() {
  if (_logPollTimeout) clearTimeout(_logPollTimeout);
  var delay = _LOG_POLL_DELAYS[Math.min(_logPollSameCount, _LOG_POLL_DELAYS.length - 1)];
  _logPollTimeout = setTimeout(pollLLMLog, delay);
}

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• api.openai.com
  • Triggering command: /usr/bin/python python -m pytest tests/ -x -q --tb=short (dns block)
  • Triggering command: /usr/bin/python python -m pytest tests/ -q --tb=short (dns block)
  • Triggering command: /usr/bin/python python -m pytest tests/ -q (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[Copilot]

Pull request overview

This PR updates the client-side LLM log polling loop to avoid permanently running fixed-interval timers by switching to a self-rescheduling setTimeout approach with an exponential-ish backoff when polls return no new entries.

Changes:

• Replaces setInterval-based polling with a _scheduleLogPoll() helper that re-arms polling via setTimeout.
• Tracks consecutive “no new entries” polls in _logPollSameCount and applies a delay from _LOG_POLL_DELAYS.
• Resets the backoff counter when new log entries are detected (and on manual clear).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

With exponential backoff, polling can reach a 10-minute delay. If a user starts a new LLM action after a long idle period, the UI may not show new /llm_log entries (and sticky status updates) for up to the full backoff delay because the backoff is only reset after a poll detects new entries. Consider resetting _logPollSameCount (and triggering an immediate/near-immediate poll) when client-side actions that generate LLM traffic start (e.g., outline generation, chapter generation/revision, illustration generation), so logs remain responsive while still backing off during true idle periods.