Survey + fixes so the pipeline passes at the Docker level (no k8s).
- Remove k8s: drop the `sandbox-k8s` job (kind/Calico/--features k8s-tests) and the
"Helm chart lints" gate step. release.yml was already k8s-clean.
- Rust job:
- `cargo fmt --all` — fix pre-existing formatting drift (fmt --check was failing).
- clippy -D warnings: fix 3 lib warnings (cm-brain sort_by_key→Reverse, cm-api
fleet.rs doc list indentation, node_rules map_or→is_none_or).
- Regenerate the .sqlx offline cache (was missing the cm-runtime run_loop test
query → offline compile failed). DB-backed tests use testcontainers at runtime.
- Set SQLX_OFFLINE=true on the rust + e2e jobs so query! macros compile against
the committed cache deterministically (no DB needed at compile time).
- Frontend job:
- Fix the 1 ESLint error (useAgentTelemetry: no setState-synchronously-in-effect;
tag the slice with agentId + derive null on mismatch).
- Fix 2 stale panel-params tests (`terminal` is a valid app id now; assert the
current APP_IDS + use a genuinely-unknown id for the reject case).
Verified locally: fmt clean, clippy --all-targets -D warnings clean (offline),
frontend lint 0 errors, tsc clean, 86/86 frontend tests pass, build OK.
Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>
352 lines
15 KiB
Rust
352 lines
15 KiB
Rust
//! Clawmates server: REST API and streaming gateway (later phases add the
|
|
//! scheduler and safety worker) composed into one binary.
|
|
|
|
mod e2e;
|
|
|
|
use std::path::{Path, PathBuf};
|
|
use std::process::ExitCode;
|
|
use std::sync::Arc;
|
|
|
|
use cm_config::{AppConfig, LlmProviderKind};
|
|
use cm_llm::{AnthropicProvider, LlmProvider, OpenAiCompatProvider, ScriptedProvider};
|
|
use cm_runtime::{Runtime, RuntimeConfig};
|
|
|
|
#[tokio::main]
|
|
async fn main() -> ExitCode {
|
|
match run().await {
|
|
Ok(()) => ExitCode::SUCCESS,
|
|
Err(message) => {
|
|
eprintln!("clawmates-server: {message}");
|
|
ExitCode::FAILURE
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Instantiates the configured LLM provider. The Anthropic key comes from
|
|
/// the environment until the secret broker lands in P2.
|
|
fn build_provider(config: &AppConfig) -> Result<Arc<dyn LlmProvider>, String> {
|
|
match config.llm.provider {
|
|
LlmProviderKind::Anthropic => {
|
|
let key = std::env::var("ANTHROPIC_API_KEY")
|
|
.map_err(|_| "llm.provider = \"anthropic\" requires ANTHROPIC_API_KEY")?;
|
|
Ok(Arc::new(AnthropicProvider::new(key)))
|
|
}
|
|
LlmProviderKind::OpenAiCompat => {
|
|
let base_url = config.llm.base_url.clone().expect("validated by cm-config");
|
|
Ok(Arc::new(OpenAiCompatProvider::new(
|
|
base_url,
|
|
std::env::var("CLAWMATES_LLM_API_KEY").ok(),
|
|
)))
|
|
}
|
|
LlmProviderKind::Scripted => {
|
|
let path = config
|
|
.llm
|
|
.scenario_path
|
|
.clone()
|
|
.expect("validated by cm-config");
|
|
let provider = ScriptedProvider::from_path(Path::new(&path))
|
|
.map_err(|e| format!("scenario load failed: {e}"))?;
|
|
Ok(Arc::new(provider))
|
|
}
|
|
}
|
|
}
|
|
|
|
/// Builds the extra named-provider registry (GLM, Kimi, …) from
|
|
/// `[[llm.providers]]`. Each is OpenAI-compatible; its key comes from the named
|
|
/// env var. A provider whose key env is unset is skipped with a warning, so a
|
|
/// missing key degrades that one selector rather than failing boot.
|
|
fn build_provider_registry(config: &AppConfig) -> cm_runtime::ProviderRegistry {
|
|
let mut map = std::collections::HashMap::new();
|
|
for p in &config.llm.providers {
|
|
match std::env::var(&p.api_key_env) {
|
|
Ok(key) if !key.is_empty() => {
|
|
let provider: Arc<dyn LlmProvider> = match p.format.as_str() {
|
|
"anthropic" => Arc::new(cm_llm::AnthropicProvider::with_base_url(
|
|
key,
|
|
p.base_url.clone(),
|
|
)),
|
|
_ => Arc::new(OpenAiCompatProvider::new(p.base_url.clone(), Some(key))),
|
|
};
|
|
map.insert(p.name.clone(), provider);
|
|
println!(
|
|
"clawmates-server: registered LLM provider '{}' ({})",
|
|
p.name, p.format
|
|
);
|
|
}
|
|
_ => eprintln!(
|
|
"clawmates-server: provider '{}' skipped — {} is unset",
|
|
p.name, p.api_key_env
|
|
),
|
|
}
|
|
}
|
|
cm_runtime::ProviderRegistry(map)
|
|
}
|
|
|
|
async fn run() -> Result<(), String> {
|
|
let config_path = PathBuf::from(
|
|
std::env::var("CLAWMATES_CONFIG").unwrap_or_else(|_| "clawmates.toml".into()),
|
|
);
|
|
let config = AppConfig::load_from(&config_path).map_err(|e| e.to_string())?;
|
|
let _telemetry = cm_telemetry::init(
|
|
"clawmates-server",
|
|
config.telemetry.otlp_endpoint.as_deref(),
|
|
)
|
|
.map_err(|e| format!("telemetry: {e}"))?;
|
|
|
|
let pool = cm_db::connect(&config.database.url, config.database.max_connections)
|
|
.await
|
|
.map_err(|e| format!("database connection failed: {e}"))?;
|
|
cm_db::MIGRATOR
|
|
.run(&pool)
|
|
.await
|
|
.map_err(|e| format!("migrations failed: {e}"))?;
|
|
|
|
if e2e::enabled() {
|
|
e2e::seed(&pool).await?;
|
|
}
|
|
|
|
// First-owner bootstrap (self-hosted local-auth installs). Provisions
|
|
// the initial workspace + Owner from env exactly once; a no-op on
|
|
// every later boot. CLAWMATES_BOOTSTRAP_OWNER_PASSWORD is the trigger.
|
|
if let Some(password) = std::env::var("CLAWMATES_BOOTSTRAP_OWNER_PASSWORD")
|
|
.ok()
|
|
.filter(|p| !p.is_empty())
|
|
{
|
|
let email = std::env::var("CLAWMATES_BOOTSTRAP_OWNER_EMAIL")
|
|
.unwrap_or_else(|_| "[email protected]".into());
|
|
let workspace = std::env::var("CLAWMATES_BOOTSTRAP_WORKSPACE")
|
|
.unwrap_or_else(|_| "My Workspace".into());
|
|
let credits = std::env::var("CLAWMATES_BOOTSTRAP_CREDITS")
|
|
.ok()
|
|
.and_then(|v| v.parse().ok())
|
|
.unwrap_or(1250);
|
|
let created = cm_auth::bootstrap_owner(&pool, &workspace, &email, &password, credits)
|
|
.await
|
|
.map_err(|e| format!("bootstrap owner: {e}"))?;
|
|
if created {
|
|
println!("clawmates-server: bootstrapped first owner {email} in '{workspace}'");
|
|
}
|
|
}
|
|
|
|
let provider = build_provider(&config)?;
|
|
let provider_registry = build_provider_registry(&config);
|
|
let blob: std::sync::Arc<dyn cm_files::BlobStore> = match config.storage.backend {
|
|
cm_config::StorageBackend::Local => std::sync::Arc::new(cm_files::LocalBlobStore::new(
|
|
PathBuf::from(&config.storage.data_dir),
|
|
)),
|
|
cm_config::StorageBackend::S3 => std::sync::Arc::new(
|
|
cm_files::S3BlobStore::connect(
|
|
config.storage.s3_endpoint.as_deref().expect("validated"),
|
|
config.storage.s3_bucket.as_deref().expect("validated"),
|
|
config.storage.s3_access_key.as_deref().unwrap_or_default(),
|
|
config.storage.s3_secret_key.as_deref().unwrap_or_default(),
|
|
)
|
|
.map_err(|e| format!("s3 storage: {e}"))?,
|
|
),
|
|
};
|
|
// The fleet node hub (live daemon channels) is shared between the sandbox
|
|
// placement provider and the API routes, so it must exist before the managers.
|
|
let node_hub = std::sync::Arc::new(cm_api::fleet::NodeHub::new());
|
|
// Environment tools need a container engine; absence is tolerated
|
|
// (shell.exec reports it per-call) so the API still serves.
|
|
let (sandboxes, browser, terminals) = if config.sandbox.enabled {
|
|
match cm_sandbox::DockerDriver::connect() {
|
|
Ok(driver) => {
|
|
let driver: std::sync::Arc<dyn cm_sandbox::SandboxDriver> =
|
|
std::sync::Arc::new(driver);
|
|
let agents = std::sync::Arc::new(
|
|
cm_runtime::SandboxManager::new(
|
|
driver.clone(),
|
|
pool.clone(),
|
|
"local",
|
|
&config.sandbox.image,
|
|
)
|
|
.with_node_provider(std::sync::Arc::new(
|
|
cm_api::fleet::HubDriverProvider::new(node_hub.clone()),
|
|
)),
|
|
);
|
|
let browser = std::sync::Arc::new(
|
|
cm_runtime::SandboxManager::new(
|
|
driver.clone(),
|
|
pool.clone(),
|
|
"local",
|
|
&config.sandbox.browser_image,
|
|
)
|
|
.with_egress(),
|
|
);
|
|
// Themed interactive terminal containers (zsh + oh-my-zsh + p10k)
|
|
// for the Terminal computer app. On the Local storage backend the
|
|
// file-drive volume is mounted (per-agent subpath) at ~/drives.
|
|
let drives = if config.storage.backend == cm_config::StorageBackend::Local {
|
|
Some(cm_runtime::DriveConfig {
|
|
volume: config.sandbox.terminal_drive_volume.clone(),
|
|
data_dir: PathBuf::from(&config.storage.data_dir),
|
|
})
|
|
} else {
|
|
None
|
|
};
|
|
let terminals = std::sync::Arc::new(
|
|
cm_runtime::TerminalManager::new(
|
|
driver,
|
|
pool.clone(),
|
|
"local",
|
|
&config.sandbox.terminal_image,
|
|
config.sandbox.terminal_egress,
|
|
drives,
|
|
)
|
|
// An agent placed on a fleet node runs its terminal there too
|
|
// (beside its sandbox), sharing the agent's node-local drives.
|
|
.with_node_provider(std::sync::Arc::new(
|
|
cm_api::fleet::HubDriverProvider::new(node_hub.clone()),
|
|
)),
|
|
);
|
|
// Boot reconciliation: any sandbox the engine still holds is an
|
|
// orphan from a dead process (we track none yet) — remove them
|
|
// before warming so a crash/redeploy can't leak containers.
|
|
agents.reconcile_orphans(std::time::Duration::ZERO).await;
|
|
browser.reconcile_orphans(std::time::Duration::ZERO).await;
|
|
terminals.reconcile_orphans().await;
|
|
let agents = if config.sandbox.warm_pool > 0 {
|
|
agents.warm(config.sandbox.warm_pool)
|
|
} else {
|
|
agents
|
|
};
|
|
// Periodic reaper: catch leaks while we run (idle > 10 min and
|
|
// owned by no live handle), checked every 5 min.
|
|
agents.clone().spawn_reaper(
|
|
std::time::Duration::from_secs(300),
|
|
std::time::Duration::from_secs(600),
|
|
);
|
|
browser.clone().spawn_reaper(
|
|
std::time::Duration::from_secs(300),
|
|
std::time::Duration::from_secs(600),
|
|
);
|
|
// Terminals: reap idle (no live session for ~2 h, so a resumable
|
|
// tmux session survives normal navigation gaps) + orphans.
|
|
terminals.clone().spawn_reaper(
|
|
std::time::Duration::from_secs(300),
|
|
std::time::Duration::from_secs(7200),
|
|
);
|
|
(Some(agents), Some(browser), Some(terminals))
|
|
}
|
|
Err(error) => {
|
|
eprintln!("clawmates-server: sandbox engine unavailable: {error}");
|
|
(None, None, None)
|
|
}
|
|
}
|
|
} else {
|
|
(None, None, None)
|
|
};
|
|
// Keep handles for the graceful-shutdown drain (SIGTERM) — the managers
|
|
// themselves are moved into the runtime config below.
|
|
let drain_agents = sandboxes.clone();
|
|
let drain_browser = browser.clone();
|
|
let drain_terminals = terminals.clone();
|
|
let runtime = Runtime::with_blob_store(
|
|
pool.clone(),
|
|
provider,
|
|
RuntimeConfig {
|
|
model: config.llm.model.clone(),
|
|
max_tokens: 4096,
|
|
broker_socket: Some(PathBuf::from(&config.broker.socket_path)),
|
|
slack_base_url: config.slack.base_url.clone(),
|
|
sandboxes,
|
|
browser,
|
|
terminals,
|
|
providers: provider_registry,
|
|
},
|
|
blob,
|
|
);
|
|
// Durable §15 path: expires overdue approvals and resumes decided runs
|
|
// even if the deciding request's process died mid-flight.
|
|
runtime.spawn_resume_sweeper(std::time::Duration::from_secs(2));
|
|
// Routine firings (§7.6).
|
|
cm_scheduler::Scheduler::new(pool.clone(), runtime.clone())
|
|
.spawn(std::time::Duration::from_secs(5));
|
|
// Durable topology run jobs: claim queued runs, drive + checkpoint per step,
|
|
// resume stale ones after a crash. Long-horizon topologies run here, not in
|
|
// the HTTP request.
|
|
cm_api::topology_worker::spawn(
|
|
pool.clone(),
|
|
runtime.clone(),
|
|
std::time::Duration::from_secs(3),
|
|
);
|
|
// Outbound-email delivery: drains the §15-gated `outbox` over SMTP. Inert
|
|
// until CLAWMATES_SMTP_* is set, so it ships safely before credentials exist.
|
|
cm_runtime::spawn_drainer(pool.clone(), std::time::Duration::from_secs(10));
|
|
// Expiry/retention sweep: expires stale auth/oauth rows and prunes old
|
|
// journal/audit rows hourly so unbounded tables don't accumulate.
|
|
cm_api::cleanup_sweeper::spawn(pool.clone(), std::time::Duration::from_secs(3600));
|
|
// Fleet backstop: a node whose heartbeats stop (without a clean channel
|
|
// close) goes offline within ~28s even if its control channel hangs.
|
|
cm_api::fleet::spawn_node_sweeper(pool.clone(), std::time::Duration::from_secs(8), 20);
|
|
// Beszel: poll each workspace's monitoring hub for rich per-node metrics.
|
|
cm_api::beszel::spawn_poller(pool.clone(), std::time::Duration::from_secs(15));
|
|
// Fleet automation: evaluate metric-threshold rules → drain/undrain/alert.
|
|
cm_api::node_rules::spawn_evaluator(pool.clone(), std::time::Duration::from_secs(20));
|
|
|
|
// Hosted identity (Clerk / OIDC): pin the issuer and load its JWKS.
|
|
let auth_verifier = match config.auth.mode {
|
|
cm_config::AuthMode::Clerk | cm_config::AuthMode::Oidc => {
|
|
let issuer = config
|
|
.auth
|
|
.issuer_url
|
|
.as_deref()
|
|
.expect("validated by config");
|
|
Some(std::sync::Arc::new(
|
|
cm_auth::JwtVerifier::discover(issuer)
|
|
.await
|
|
.map_err(|e| format!("identity issuer: {e}"))?,
|
|
))
|
|
}
|
|
cm_config::AuthMode::Local => None,
|
|
};
|
|
|
|
let mut app = cm_api::router(
|
|
cm_api::AppState::new(pool, runtime)
|
|
.with_node_hub(node_hub.clone())
|
|
.with_broker(PathBuf::from(&config.broker.socket_path))
|
|
.with_oauth(config.oauth.clone())
|
|
.with_billing(config.billing.clone())
|
|
.with_file_root(
|
|
(config.storage.backend == cm_config::StorageBackend::Local)
|
|
.then(|| PathBuf::from(&config.storage.data_dir)),
|
|
)
|
|
.with_per_signup_workspace(config.auth.per_signup_workspace)
|
|
.pipe_auth_verifier(auth_verifier),
|
|
);
|
|
if e2e::enabled() {
|
|
app = app.merge(e2e::slack_sink_router());
|
|
}
|
|
let listener = tokio::net::TcpListener::bind(config.listen_addr)
|
|
.await
|
|
.map_err(|e| format!("bind {} failed: {e}", config.listen_addr))?;
|
|
println!("clawmates-server listening on {}", config.listen_addr);
|
|
// Graceful shutdown: on SIGTERM/Ctrl-C, stop accepting, finish in-flight
|
|
// requests, then DRAIN the sandbox managers so no container is left running.
|
|
let shutdown = async move {
|
|
let mut term = tokio::signal::unix::signal(tokio::signal::unix::SignalKind::terminate())
|
|
.expect("install SIGTERM handler");
|
|
tokio::select! {
|
|
_ = term.recv() => {}
|
|
_ = tokio::signal::ctrl_c() => {}
|
|
}
|
|
eprintln!("clawmates-server: shutdown signal received");
|
|
};
|
|
axum::serve(listener, app)
|
|
.with_graceful_shutdown(shutdown)
|
|
.await
|
|
.map_err(|e| format!("server error: {e}"))?;
|
|
eprintln!("clawmates-server: draining sandboxes…");
|
|
if let Some(agents) = drain_agents {
|
|
agents.shutdown().await;
|
|
}
|
|
if let Some(browser) = drain_browser {
|
|
browser.shutdown().await;
|
|
}
|
|
if let Some(terminals) = drain_terminals {
|
|
terminals.shutdown().await;
|
|
}
|
|
Ok(())
|
|
}
|