This PR does three things:
- Exports derp expvars to the pprof endpoint
- Exports the expvar metrics as prometheus metrics in both coderd and
wsproxy
- Updates our tailscale to a fix I also had to make to avoid a data race
condition
I generated this with mux but I also manually tested that the metrics
were getting properly emitted
- Previously all tests were sharing the global http.Transport meaning on
`Close` it would close connections presumed to be idle for other tests.
fixes https://github.com/coder/internal/issues/112
The provisioner state for a workspace build was being loaded for every
long-lived agent rpc connection. Since this state can be anywhere from
kilobytes to megabytes this can gradually cause the `coderd` memory
footprint to grow over time. It's also a lot of unnecessary allocations
for every query that fetches a workspace build since only a few callers
ever actually reference the provisioner state.
This PR removes it from the returned workspace build and adds a query to
fetch the provisioner state explicitly.
Adds coderd_template_workspace_build_duration_seconds histogram that
tracks the full duration from workspace build creation to agent ready.
This captures the complete user-perceived build time including
provisioning and agent startup.
The metric is emitted when the agent reports ready/error/timeout via the
lifecycle API, ensuring each build is counted exactly once per replica.
Adds additional logs for determining what signal the agent receives
prior to shut down. Also helps distinguish whether the signal originated
at the agent or reaper.
Adds support for filtering workspaces by health status using
healthy:true or healthy:false in the search query.
This is done by changing `has-agent` to accept a list of statuses and
aliasing `health:true` to `has-agent:connected` and `healthy:false` to
`has-agent:timeout,disconnected`.
Fixes#21623
The reaper (PID 1) now returns the child's exit code instead of always
exiting 0. Signal termination uses the standard Unix convention of 128 +
signal number.
fixes#21661
I'll be honest I'm not even really sure the point of this test but it
was failing due to
```
2025-06-16T15:01:54.0863251Z Error: Received unexpected error:
2025-06-16T15:01:54.0863554Z acquire job:
2025-06-16T15:01:54.0864230Z github.com/coder/coder/v2/coderd/provisionerdserver.(*server).AcquireJob
2025-06-16T15:01:54.0865173Z /home/runner/work/coder/coder/coderd/provisionerdserver/provisionerdserver.go:329
2025-06-16T15:01:54.0865683Z - failed to acquire job:
2025-06-16T15:01:54.0866374Z github.com/coder/coder/v2/coderd/provisionerdserver.(*Acquirer).AcquireJob
2025-06-16T15:01:54.0867262Z /home/runner/work/coder/coder/coderd/provisionerdserver/acquirer.go:148
2025-06-16T15:01:54.0867819Z - pq: canceling statement due to user request
```
which is certainly unintended.
- Removes displaying XRay scan results in the dashboard. I'm not sure
anyone was even using this integration so it's just debt for us to
maintain. We can open up a separate issue to get rid of the db tables
once we know for sure that we haven't broken anyone.
- Update go.mod to use Go 1.24.1
- Update GitHub Actions setup-go action to use Go 1.24.1
- Fix linting issues with golangci-lint by:
- Updating to golangci-lint v1.57.1 (more compatible with Go 1.24.1)
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
---------
Co-authored-by: Claude <claude@anthropic.com>
- Fixes an issue where an init function
causes writes to the terminal when using `agent-exec` which results in raw
ansi characters being printed to the web terminal.
- Was also causing significant latency in launching a web terminal
- Filters env vars specific to agent-exec from the exec'd process. This
is to prevent any issues when developing Coder in Coder, particularly
agent tests in the cli pkg.
- Fixes an issue where oom scores would fail to be adjusted in cases
where the `coder` binary has capabilities set on it. This is because
`PR_SET_DUMPABLE` is set to `0` when a process is executed with elevated
capabilities. The fix is to flip `PR_SET_DUMPABLE` to `1` prior to
writing to `oom_score_adj`.
- Integrates the `agentexec` pkg into the agent and removes the
legacy system of iterating over the process tree. It adds some linting
rules to hopefully catch future improper uses of `exec.Command` in the package.
- We were instantiating a cryptokey cache with a vanilla reference to
the database instead of one wrapped by dbcrypt.
- Fixes an issue where failing to instantiate unrelated keycaches does
not fatally error out.
- Fixes an issue where building the Docker image failed due to moving
the directory hosting the Dockerfile
- Removed the Palo Alto scanning since our subscription there is set to
expire. Trivy is still running though.
- Adds a `jwtutils` package to be shared amongst the various
packages in the codebase that make use of JWTs. It's intended to help us
standardize on one library instead of some implementations using
`go-jose` and others using `golang-jwt`.
The main reason we're converging on `go-jose` is due to its support for
JWEs, `golang-jwt` also has a repo to handle it but it doesn't look
maintained: https://github.com/golang-jwt/jwe
- Adds the database implementation for fetching and caching keys
used for JWT signing. It's been merged into the `keyrotate` pkg and
renamed to `cryptokeys` since they're coupled concepts.
* fix: avoid deleting peers on graceful close
- Fixes an issue where a coordinator deletes all
its peers on shutdown. This can cause disconnects
whenever a coderd is redeployed.
- It's bafflingly buggy and is a source of annoyance for virtually the
whole team.
- Will revisit if we don't have alternatives to catching invalid licenses.
Jon Ayers