* feat: allow scaling down scaletest environments
* fix bugged namespace deletion
* misc fixes to scaletest.sh
* destroy namespaces is a no-op as the cluster will be gone anyway
Updates scaletest terraform with learnings from #8213:
- Increase max pods per node to 256
- Decrease CPU requests for test workspace to allow maxing out workspaces per node
- Explicitly set CODER_ACCESS_URL for ssh to work
- Explicitly disable rate limits in coderd
- Increase DB size for medium and large scenarios
- Mount cache volume directly under /tmp/coder instead of /tmp.
- Plumb through requests and limits for workspaces
- Plumb through requests for coderd
* chore: scaletest: plumb through more options
* bump terraform version
* scaletest.sh: pprof during traffic gen
* cli/scaletest: actually wait for prometheus metrics to be scraped
* increase prometheus wait
- Adds a convenience script `scaletest.sh` to automate process of running scale tests
- Enables pprof endpoint by default, and captures pprof traces before tearing down infra.
- Improves idempotency of coder_init.sh
- Removes the promtest.Float64 invocations in workspacetraffic runner, these metrics will be in prometheus.
- Increases default workspace traffic output to 40KB/s/workspace.
- Exposes reads/writes from scaletest traffic generation (default: 0.0.0.0:21112)
- Adds self-hosted prometheus with remote_write to loadtest terraform
- Adds convenience script to run a traffic generation test
This PR adds a scaletest workspace-traffic command for load testing. This opens a
ReconnectingPTY connection to each scaletest workspace (via coderd) and
concurrently writes and reads random data to/from the PTY. Payloads are of the
form #${RANDOM_ALPHANUMERIC_STRING}, which essentially drops garbage
comments in the remote shell, and should not result in any commands being executed.
* fix: use a waitgroup to ensure all connections are cleaned up in agent
There was a race where connections would be created at the same time as close.
The `net.Conn` produced by Tailscale doesn't close then the listener does.
* Remove accidental test
* chore: rename `AgentConn` to `WorkspaceAgentConn`
The codersdk was becoming bloated with consts for the workspace
agent that made no sense to a reader. `Tailnet*` is an example
of these consts.
* chore: remove `Get` prefix from *Client functions
* chore: remove `BypassRatelimits` option in `codersdk.Client`
It feels wrong to have this as a direct option because it's so infrequently
needed by API callers. It's better to directly modify headers in the two
places that we actually use it.
* Merge `appearance.go` and `buildinfo.go` into `deployment.go`
* Merge `experiments.go` and `features.go` into `deployment.go`
* Fix `make gen` referencing old type names
* Merge `error.go` into `client.go`
`codersdk.Response` lived in `error.go`, which is wrong.
* chore: refactor workspace agent functions into agentsdk
It was odd conflating the codersdk that clients should use
with functions that only the agent should use. This separates
them into two SDKs that are closely coupled, but separate.
* Merge `insights.go` into `deployment.go`
* Merge `organizationmember.go` into `organizations.go`
* Merge `quota.go` into `workspaces.go`
* Rename `sse.go` to `serversentevents.go`
* Rename `codersdk.WorkspaceAppHostResponse` to `codersdk.AppHostResponse`
* Format `.vscode/settings.json`
* Fix outdated naming in `api.ts`
* Fix app host response
* Fix unsupported type
* Fix imported type
* Add extio
* feat: Add `vscodeipc` subcommand for VS Code Extension
This enables the VS Code extension to communicate with a Coder client.
The extension will download the slim binary from `/bin/*` for the
respective client architecture and OS, then execute `coder vscodeipc`
for the connecting workspace.
* Add authentication header, improve comments, and add tests for the CLI
* Update cli/vscodeipc_test.go
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* Update cli/vscodeipc_test.go
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* Update cli/vscodeipc/vscodeipc_test.go
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
* Fix requested changes
* Fix IPC tests
* Fix shell execution
* Fix nix flake
* Silence usage
Co-authored-by: Mathias Fredriksson <mafredri@gmail.com>
They way the reconnectingpty tests behave inherently will cause the
agent to occasionally log an error (e.g. due to test disconnecting at a
certain time), allowing these error logs to fail the test will cause
these tests to be flakey.
It's best for these tests to only rely on the observed behavior.
Cian Johnston