docs: update Tallyman Agent Time reporting - #28275
Conversation
Moves Coder Agents usage reporting into the Licensing & Usage page and updates the documentation to describe Agent Time and the hb_agent_runtime_v1 Tallyman payload. Removes the superseded Usage Data Reporting page and its manifest entry, and removes the obsolete AI Governance link to that page.
Removes the Coder Tasks and Agent Workspace Build usage-measurement section from the AI Governance page. AI Governance is not a prerequisite for usage-based billing, so this content doesn't belong on the governance page.
Rephrases the Agent Time definition to avoid the word runtime, and removes the stream-retries, local-tool-execution, and failed-model-call bullets per review feedback.
|
I'll open a PR in coder/coder.com to redirect the removed page. |
Documentation CheckThis is a docs PR, so the findings below are about consistency gaps the change leaves behind, not missing docs for code. Updates Needed
New Documentation Needed
External follow-up
Automated review via Coder Agents |
…ce cleanup Update glossary, tasks, and index pages to stop pointing readers to ai-governance.md for Agent Workspace Build details that were removed. Delete the orphaned ai-governance-awb-usage.png screenshot.
|
Replying to the doc-check comment. All items addressed in commit a2c5304:
PR generated with Coder Agents |
…ks.md Remove the Agent Workspace Build and Coder Tasks glossary entries and the Coder Tasks mention from the Coding agent entry. Revert the tasks.md note to its original text since the Coder Tasks page removal is handled by a separate PR.
|
Updated approach per feedback (commits a2c5304 and 5dfa2ad):
PR generated with Coder Agents |
Coder Agents is a native harness, not a runner for third-party agents like Claude Code. The Coding agent entry now defines the term without tying it to the Coder Agents product.
Auditing, Restricting, Centralizing, Standardizing, and Measuring headings now use imperative mood per the style guide.
|
Two accuracy tweaks for the Agent Time usage reporting section (
Suggested wording for the cadence paragraph:
|
Clarify context compaction as summarization and describe interrupted generation as time streamed or spent executing tools before the interrupt, kept on the partial assistant messages.
…xcludes Add abbreviated bullets for failed/retried model calls and client-executed tools handed off to external coding agents, per review feedback.
Distinguish the hourly event generation cadence from the 17-minute publish check interval, and document the 7-day backfill window with its 100-event batch size.
applied rec'd changes. can add follow-ups in another PR if/when necessary.

Moves Coder Agents usage reporting into the Licensing & Usage page and updates the documentation to describe Agent Time and the
hb_agent_runtime_v1Tallyman payload.Removes the superseded Usage Data Reporting page and its manifest entry, and removes the obsolete AI Governance link to that page.
PR generated with Coder Agents