fix: align OpenAI cache affinity and use uuidv7 session ids

packages/coding-agent/CHANGELOG.md

@@ -55,6 +55,7 @@ How to disable it:
 - Fixed Gemma 4 thinking level mapping to route between `MINIMAL` and `HIGH`, and map Pi reasoning levels to the model's supported thinking levels ([#2903](https://github.com/badlogic/pi-mono/pull/2903) by [@aadishv](https://github.com/aadishv))
 - Fixed Gemini 2.5 Flash Lite minimal thinking budget to use the model's supported 512-token minimum instead of the regular Flash 128-token minimum, avoiding invalid thinking budget errors ([#2861](https://github.com/badlogic/pi-mono/pull/2861) by [@JasonOA888](https://github.com/JasonOA888))
 - Fixed OpenAI Codex Responses requests to forward configured `serviceTier` values, restoring service-tier selection for Codex sessions ([#2996](https://github.com/badlogic/pi-mono/pull/2996) by [@markusylisiurunen](https://github.com/markusylisiurunen))
+- Fixed newly generated session IDs to use UUIDv7, improving time locality for session-based request routing ([#3018](https://github.com/badlogic/pi-mono/pull/3018) by [@steipete](https://github.com/steipete))
 - Fixed `Container.render()` stack overflow on long sessions by replacing `Array.push(...spread)` with a loop-based push, preventing `RangeError: Maximum call stack size exceeded` when child output exceeds the V8 call stack argument limit ([#2651](https://github.com/badlogic/pi-mono/issues/2651))
 - Fixed editor sticky-column tracking around paste markers so vertical cursor navigation restores the column from before the cursor entered a paste marker instead of jumping inside or past pasted content ([#3092](https://github.com/badlogic/pi-mono/pull/3092) by [@Perlence](https://github.com/Perlence))
 - Fixed queued messages typed during `/tree` branch summarization to flush automatically after navigation completes, so they no longer remain stuck in the steering queue ([#3091](https://github.com/badlogic/pi-mono/pull/3091) by [@Perlence](https://github.com/Perlence))