> 🤖 This PR was written by Coder Agents on behalf of Jake Howell.
Closes
[DEVEX-381](https://linear.app/codercom/issue/DEVEX-381/flake-test-tasksendwaitsforworkingappstate).
Follow-up to #25648 and #25858, which addressed a different symptom of
the same test.
## Symptom
```
task_send_test.go:348: context expired while waiting for trap: context deadline exceeded
--- FAIL: Test_TaskSend/WaitsForWorkingAppState (26.02s)
```
Windows-only, on `test-go-pg (windows-2022)`. Reported four times since
#25648 landed (2026-06-02, 2026-06-10, 2026-07-01).
## Root cause
The test:
1. `setupCLITaskTest` inserts `workspace_app_status(state=idle)` at the
end of setup.
2. `WaitsForWorkingAppState` then inserts
`workspace_app_status(state=working)` before starting the CLI.
3. Both are persisted via `dbtime.Now()`, which rounds to microseconds.
Windows `time.Now()` resolution is coarser than that (often ~1 ms or
worse), so back-to-back calls frequently round to the same microsecond.
4. `GetLatestWorkspaceAppStatusesByWorkspaceIDs` has no tiebreaker:
```sql
ORDER BY workspace_id, created_at DESC
```
Its sibling `GetLatestWorkspaceAppStatusByAppID` already uses `ORDER BY
created_at DESC, id DESC` for exactly this reason. When the two rows
collide, Postgres picks either.
5. On the failing runs, the query returned the `idle` row.
`waitForTaskIdle` saw idle on the first poll, returned nil, `TaskSend`
proceeded, and the CLI completed successfully in ~5 s.
6. But the test was blocked at `resetTrap.MustWait(ctx)` waiting for a
**second** `ticker.Reset` that never happened. `WaitLong = 25s` elapsed,
line 348 failed.
CI log confirms the sequence: only one `Ticker.Reset(5s)` is caught,
then `Ticker.Stop([]) call, matched 0 traps` (from `defer
ticker.Stop()`), then the trap wait times out.
This is the same class of flake Spike documented in #15923 and #21332
("Windows in particular doesn't have high-resolution timers"), just
hidden behind a SQL `ORDER BY`.
## Fix
Two changes:
1. **`coderd/database/queries/workspaceapps.sql`**: add an `id DESC`
tiebreaker to `GetLatestWorkspaceAppStatusesByWorkspaceIDs`, matching
`GetLatestWorkspaceAppStatusByAppID`. Makes the query deterministic when
`created_at` collides.
2. **`cli/task_test.go` / `cli/task_send_test.go`**: add a
`withoutInitialAppStatus()` option to `setupCLITaskTest` and use it from
`WaitsForWorkingAppState`. The test now inserts a single `working` row,
so the collision cannot happen in the first place. Belt-and-braces with
change 1.
Comments in both places reference DEVEX-381 and #21332 so the next agent
doesn't have to re-derive this.
## Verification
- `go test ./cli -run 'Test_TaskSend' -count=1`: all 12 subtests pass,
`WaitsForWorkingAppState` completes in ~5.6 s (was ~16 s previously due
to a longer poll loop).
- Stress: 20 sequential runs of `WaitsForWorkingAppState` on Linux,
race-enabled binary, all pass in ~5.5 s each.
- `go test ./coderd -run 'AppStatus|Task' -count=1` passes.
- `go vet ./coderd/database/... ./cli/...` clean.
- `make lint/emdash` clean.
- `gofmt` clean.
Not reproducible on Linux (real time between the two patches is orders
of magnitude larger than microsecond); the Windows path is fixed by
making the ordering deterministic and by not creating the collision in
the first place.
<details>
<summary>Implementation plan & decision log</summary>
### Investigation
1. Pulled the failing job log for run `28483879823/job/84428355669`.
2. Traced the mock-clock trap sequence: one `NewTicker` and exactly one
`Ticker.Reset(5s)` were caught, then `Ticker.Stop([]) call, matched 0
traps` fires (the `defer ticker.Stop()` on `waitForTaskIdle` return).
This proves `waitForTaskIdle` returned after a single poll, not that the
trap machinery hung.
3. The command exited with `<nil>` (`clitest.go:299: command "coder task
send" exited with error: <nil>`) and a `POST /send` completed in 5.4 s.
So the CLI succeeded; the test's own trap wait is what timed out.
4. The only `waitForTaskIdle` return-nil paths are `Active +
CurrentState.State in {Idle, Complete, Failed}` and `Active +
CurrentState == nil past 30s grace`. First observation of nil cannot be
past 30s. So `TaskByID` must have returned `State == Idle`.
5. Traced `TaskByID` → `taskGet` → `workspaceData` →
`GetLatestWorkspaceAppStatusesByWorkspaceIDs`. Found the missing
tiebreaker; the sibling query one line above
(`GetLatestWorkspaceAppStatusByAppID`) already had it.
6. Confirmed the two `PATCH /app-status` calls in the Windows log
happened at `00:26:13.077` and `00:26:13.093`, well within Windows timer
resolution.
7. Confirmed `dbtime.Now()` rounds to microseconds; Windows `time.Now()`
doesn't have that precision, so `Round(time.Microsecond)` on two calls
close together frequently produces equal values.
### Prior art from Spike
- #15923: loosened `HeartbeatPeriod * 9/10` to `3/4` for Windows.
- #21332: switched `assert.After` to `assert.NotBefore` because
timestamps can equal on Windows.
Both explicitly cite "Windows doesn't always have high-resolution timers
available."
### Considered alternatives
- **Only fix the test.** Works today but leaves the SQL query
non-deterministic; another test that relies on
`GetLatestWorkspaceAppStatusesByWorkspaceIDs` could hit the same
collision.
- **Only fix the SQL query.** Would give a stable answer but not
necessarily the *right* one. If both patches share a `created_at`, `id
DESC` picks whichever UUID sorted higher, still random with respect to
insertion order.
- **Make `dbtime.Now()` monotonic per process.** Cleanest at the source,
but affects every timestamp in the database and has broader implications
than a targeted flake fix.
Going with both the query fix (defense in depth, matches existing
pattern) and the test fix (eliminates the collision at the source) is
the smallest change that closes the flake and hardens the query.
### Rejected commit-message scopes
Changes touch both `cli/` and `coderd/database/`, so per AGENTS.md the
scope is omitted for the cross-cutting commit and PR title.
</details>
Relates to https://github.com/coder/internal/issues/1259
Adds new database queries and telemetry collection functions to gather
task lifecycle events (pause/resume cycles, idle time) for analytics.
Task events track pause/resume activity, idle duration before pausing,
paused duration, and time from resume to first app status, filtered to
recent activity based on the telemetry snapshot interval.
🤖 Created with Mux (Opus 4.6).
## Description
This PR fixes an issue where `GetLatestWorkspaceAppStatusesByAppID`
returned an unbounded number of rows for a given app ID, which could
cause performance issues for noisy or long-running AI tasks.
## Impact
This change reduces database query overhead for workspace app status
updates, particularly for busy AI tasks that update their status
frequently. Previously, fetching the latest status would return all
historical statuses, now it returns only the most recent one.
Fixes#20862
---
🤖 This change was written by Claude Sonnet 4.5 Thinking using [mux](https://github.com/coder/mux) and reviewed by a human 🏄🏻♂️
## Description
Send a notification to the workspace owner when an AI task’s app state
becomes `Working` or `Idle`.
An AI task is identified by a workspace build with `HasAITask = true`
and `AITaskSidebarAppID` matching the agent app’s ID.
## Changes
* Add `TemplateTaskWorking` notification template.
* Add `TemplateTaskIdle` notification template.
* Add `GetLatestWorkspaceAppStatusesByAppID` SQL query to get the
workspace app statuses ordered by latest first.
* Update `PATCH /workspaceagents/me/app-status` to enqueue:
* `TemplateTaskWorking` when state transitions to `working`
* `TemplateTaskIdle` when state transitions to `idle`
* Notification labels include:
* `task`: task initial prompt
* `workspace`: workspace name
* Notification dedupe: include a minute-bucketed timestamp (UTC
truncated to the minute) in the enqueue data to allow identical content
to resend within the same day (but not more than once per minute).
Closes: https://github.com/coder/coder/issues/19776
https://github.com/coder/coder/pull/17163 introduced the
`workspace_app_statuses` table. Two of these fields
(`needs_user_attention`, `icon`) turned out to be surplus to
requirements.
- Removes columns `needs_user_attention` and `icon` from
`workspace_app_statuses`
- Marks the corresponding fields of `codersdk.WorkspaceAppStatus` as
deprecated.
This does ~95% of the backend work required to integrate the AI work.
Most left to integrate from the tasks branch is just frontend, which
will be a lot smaller I believe.
The real difference between this branch and that one is the abstraction
-- this now attaches statuses to apps, and returns the latest status
reported as part of a workspace.
This change enables us to have a similar UX to in the tasks branch, but
for agents other than Claude Code as well. Any app can report status
now.
* feat: use app wildcards for apps if configured
* feat: relative_path -> subdomain
- rename relative_path -> subdomain when referring to apps
- migrate workspace_apps.relative_path to workspace_apps.subdomain
- upgrade coder/coder terraform module to 0.5.0
* feat: Add anonymized telemetry to report product usage
This adds a background service to report telemetry to a Coder
server for usage data. There will be realtime event data sent
in the future, but for now usage will report on a CRON.
* Fix flake and requested changes
* Add reporting options for setup
* Add reporting for workspaces
* Add resources as they are reported
* Track API key usage
* Ensure telemetry is tracked prior to exit
* feat: Add app support
This adds apps as a property to a workspace agent.
The resource is added to the Terraform provider here:
https://github.com/coder/terraform-provider-coder/pull/17
Apps will be opened in the dashboard or via the CLI
with `coder open <name>`. If `command` is specified, a
terminal will appear locally and in the web. If `target`
is specified, the browser will open to an exposed instance
of that target.
* Compare fields in apps test
* Update Terraform provider to use relative path
* Add some basic structure for routing
* chore: Remove interface from coderd and lift API surface
Abstracting coderd into an interface added misdirection because
the interface was never intended to be fulfilled outside of a single
implementation.
This lifts the abstraction, and attaches all handlers to a root struct
named `*coderd.API`.
* Add basic proxy logic
* Add proxying based on path
* Add app proxying for wildcards
* Add wsconncache
* fix: Race when writing to a closed pipe
This is such an intermittent race it's difficult to track,
but regardless this is an improvement to the code.
* fix: Race when writing to a closed pipe
This is such an intermittent race it's difficult to track,
but regardless this is an improvement to the code.
* fix: Race when writing to a closed pipe
This is such an intermittent race it's difficult to track,
but regardless this is an improvement to the code.
* fix: Race when writing to a closed pipe
This is such an intermittent race it's difficult to track,
but regardless this is an improvement to the code.
* Add workspace route proxying endpoint
- Makes the workspace conn cache concurrency-safe
- Reduces unnecessary open checks in `peer.Channel`
- Fixes the use of a temporary context when dialing a workspace agent
* Add embed errors
* chore: Refactor site to improve testing
It was difficult to develop this package due to the
embed build tag being mandatory on the tests. The logic
to test doesn't require any embedded files.
* Add test for error handler
* Remove unused access url
* Add RBAC tests
* Fix dial agent syntax
* Fix linting errors
* Fix gen
* Fix icon required
* Adjust migration number
* Fix proxy error status code
* Fix empty db lookup
Jake Howell