root/goodbuddy
1
0
Fork 0
Code Issues Pull Requests Actions 3 Packages Projects Releases Wiki Activity

perf: track execution duration by state and bound summary I/O

Browse Source
This commit is contained in:
mesalogo
2026-10-04 16:21:55 +08:00
parent 2c2b6c31eb
commit 4d18794e74
31 changed files with 987 additions and 1397 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/architecture/runtime-storage.md
+9 -17
View File
@@ -10,20 +10,11 @@ recovery. `SubagentProgressStorage` caches event encoding state, not task statis
## Runtime Read Design
Execution statistics use one lazily started worker with a read-only connection.
The worker performs the existing evidence query inside a short read transaction.
Main supplies the current reply lease IDs and receives only scalar summaries.
There is no worker pool or duplicate persisted summary. The worker is terminated
when its owning database closes; errors reject reads, without a synchronous Main
fallback. In-memory test databases use the synchronous query because independent
connections cannot access the same private in-memory database.
Derived caches must specify scope, capacity, lifetime and invalidation. Statistics
keep at most eight scopes for 30 seconds. SQLite `data_version` detects other
connections' commits; `total_changes()` detects writes on the owning connection.
Reads inside an existing transaction clear and bypass the cache, so rolled-back
values cannot escape. Closing the connection clears its cache. Live leases bypass
the cache because elapsed time changes without a database write.
Execution duration is persisted on running-state transitions in Main. Snapshots
read small indexed timing rows; they require no historical evidence query, worker
or result cache. Main owns current segments independently of renderer visibility.
The [duration design](../features/task-and-job/technical-design.md) defines the
state boundaries, restart handling and incomplete history.
Read transactions end before replies are posted. Long-lived read transactions
would prevent WAL checkpoints from reclaiming old pages. Each reader has its own
@@ -36,9 +27,10 @@ rules, without introducing a general query protocol before a second use case exi
Hidden task panels use React `Activity` to preserve state while scheduling hidden
render work below visible updates. Panel factories run inside that boundary.
Terminal and browser panels retain their existing mounted effects and sessions.
Statistics refresh only when the task center and document are visible. Results
must remain scoped and late responses from a previous selection must be discarded.
Duration snapshots refresh on Main change notifications while the task center is
open. The renderer ticks locally for open segments and fetches fresh snapshots on
reopening. Results remain scoped; superseded responses are discarded.
The [task statistics design](../features/task-and-job/technical-design.md) owns the
query contract and evidence rules. Tests and timings demonstrate only their fixture
snapshot contract and timing rules. Tests and timings demonstrate only their fixture
conditions; they do not establish a latency guarantee for arbitrary retained history.