smartbotic/src/webserver/webserver_service.hpp

#pragma once

#include <memory>
#include <mutex>
#include <set>
#include "http_server.hpp"
#include "websocket_server.hpp"
#include "auth/jwt_utils.hpp"
#include "auth/auth_store.hpp"
#include "auth/auth_middleware.hpp"
#include "runners/runner_registry.hpp"
#include "runners/load_balancer.hpp"
#include "storage/storage_client.hpp"
#include "config/config_loader.hpp"
#include "scheduler/workflow_scheduler.hpp"
#include "scheduler/database_watcher.hpp"
#include "api/project_controller.hpp"
#include "auth/access.hpp"
#include "retention/retention_service.hpp"
#include "settings/settings_store.hpp"
#include "settings/settings_accessor.hpp"

namespace smartbotic::webserver::nodes {
    class NodeStore;
}

namespace smartbotic::webserver::grpc {
    class NodeSyncServer;
    class CredentialServer;
    class WorkflowControlServiceImpl;
}

namespace smartbotic::credentials {
    class CredentialStore;
}

namespace smartbotic::webserver::api {
    class AuthController;
    class UserController;
    class WorkflowController;
    class WorkflowGroupController;
    class ExecutionController;
    class FileController;
    class ProxyController;
    class NodeController;
    class RunnerController;
    class WebhookController;
    class DatabaseController;
    class CredentialController;
    class CertificateController;
    class SettingsController;
}

namespace smartbotic::webserver {

// Server-side projection of the executions collection. The list endpoint reads
// this instead of the collection so the database never sends nodeExecutions,
// which carry base64 images and dwarf every other field.
inline constexpr const char* kExecutionsSummaryView = "executions_summary";

struct CredentialsConfig {
    std::string master_key = "dev-key-change-in-production";
    int pbkdf2_iterations = 100000;
};

struct WebServerServiceConfig {
    int http_port = 8080;
    int node_sync_port = 9002;      // gRPC port for node sync service
    int credential_service_port = 9003;  // gRPC port for credential service
    std::string static_files_path = "./webui/dist";
    std::string database_address = "localhost:9004";
    std::string database_project = "smartbotic-automation";
    int max_upload_mb = 32;
    // Bounds WebhookController's dispatch pool, which runs immediate-mode
    // form submissions off the request thread. The form endpoint is public
    // and unauthenticated by design, so these keep a burst of submissions
    // from spawning without limit.
    int form_dispatch_threads = 4;
    int form_dispatch_queue_capacity = 32;
    // Bounds WorkflowController's dispatch pool, which runs the editor's
    // "Execute" button off the request thread for the same reason the form
    // pool above exists - the runner's ExecuteWorkflow RPC is synchronous
    // regardless of wait_for_completion. This path is authenticated, so the
    // exhaustion risk is lower than the form pool's, but bounding it still
    // costs nothing.
    int workflow_dispatch_threads = 4;
    int workflow_dispatch_queue_capacity = 32;
    runners::RunnerRegistryConfig runner_config;
    runners::LoadBalancerConfig load_balancer_config;
    auth::JwtUtils::Config jwt_config;
    CredentialsConfig credentials_config;
};

class WebServerService {
public:
    explicit WebServerService(const WebServerServiceConfig& config);
    ~WebServerService();

    static WebServerServiceConfig loadConfig(const std::filesystem::path& path);

    void start();
    void stop();

    // Access components
    auth::AuthStore& authStore() { return *auth_store_; }
    runners::RunnerRegistry& runnerRegistry() { return *runner_registry_; }
    WebSocketServer& wsServer() { return *ws_server_; }
    WorkflowScheduler& scheduler() { return *scheduler_; }

private:
    void setupRoutes();
    void loadScheduledWorkflows();
    void ensureExecutionsSummaryView();
    void executeScheduledWorkflow(const std::string& workflow_id,
                                  const std::string& trigger_node_id,
                                  const std::string& trigger_type,
                                  const nlohmann::json& extra_trigger_data = nlohmann::json::object());

    // Runs the workflow a failing workflow nominates as its error handler, if
    // it names one. Called when an execution reports failure.
    // Count how many times in a row a workflow has failed, and switch it off
    // when that passes what its settings allow. The count lives on the workflow
    // rather than in memory: a restart must not forgive a workflow that has
    // been failing for an hour, and the number is worth being able to see.
    void noteExecutionOutcome(const std::string& workflow_id,
                              const std::string& execution_id,
                              bool failed,
                              const std::string& error);

    // Close executions a runner cannot possibly still be running. A runner
    // holds what it is executing in memory, so anything recorded as running
    // against one that has just come up has nobody left to finish it - and
    // would sit at "running" until somebody noticed and pressed Stop.
    void reconcileOrphanedExecutions(const std::string& runner_id,
                                     const std::string& address);

    void runErrorWorkflow(const std::string& failed_workflow_id,
                          const std::string& failed_execution_id,
                          const std::string& error_message);

    WebServerServiceConfig config_;

    // Core components
    std::unique_ptr<storage::StorageClient> storage_;
    // Built right after storage_ and before anything that reads a live
    // setting - a live accessor handed out before this exists would be a
    // reference to nothing.
    std::unique_ptr<settings::SettingsStore> settings_store_;
    std::unique_ptr<settings::SettingsAccessor> settings_;
    std::unique_ptr<auth::JwtUtils> jwt_;
    std::unique_ptr<auth::AuthStore> auth_store_;
    std::unique_ptr<auth::AuthMiddleware> auth_middleware_;
    std::unique_ptr<runners::RunnerRegistry> runner_registry_;
    std::unique_ptr<runners::LoadBalancer> load_balancer_;
    std::unique_ptr<nodes::NodeStore> node_store_;
    std::unique_ptr<retention::RetentionService> retention_;
    std::unique_ptr<credentials::CredentialStore> credential_store_;

    // Servers
    std::unique_ptr<HttpServer> http_server_;
    std::unique_ptr<WebSocketServer> ws_server_;
    std::unique_ptr<grpc::NodeSyncServer> node_sync_server_;
    std::unique_ptr<grpc::CredentialServer> credential_server_;
    std::unique_ptr<grpc::WorkflowControlServiceImpl> workflow_control_service_;

    // Scheduler
    std::unique_ptr<WorkflowScheduler> scheduler_;
    std::unique_ptr<DatabaseWatcher> db_watcher_;
    std::unique_ptr<auth::AccessControl> access_;
    std::unique_ptr<api::ProjectController> project_ctrl_;

    // Controllers (must outlive httplib::Server callbacks)
    std::unique_ptr<api::AuthController> auth_ctrl_;
    std::unique_ptr<api::UserController> user_ctrl_;
    std::unique_ptr<api::WorkflowController> workflow_ctrl_;
    std::unique_ptr<api::WorkflowGroupController> workflow_group_ctrl_;
    std::unique_ptr<api::ExecutionController> execution_ctrl_;
    std::unique_ptr<api::FileController> file_ctrl_;
    std::unique_ptr<api::ProxyController> proxy_ctrl_;

    // Execution events can arrive more than once for the same execution, so a
    // failure is remembered briefly to keep one failure from starting the error
    // workflow twice.
    std::mutex handled_failures_mutex_;
    std::set<std::string> handled_failures_;
    std::unique_ptr<api::NodeController> node_ctrl_;
    std::unique_ptr<api::RunnerController> runner_ctrl_;
    std::unique_ptr<api::WebhookController> webhook_ctrl_;
    std::unique_ptr<api::DatabaseController> database_ctrl_;
    std::unique_ptr<api::CredentialController> credential_ctrl_;
    std::unique_ptr<api::CertificateController> certificate_ctrl_;
    std::unique_ptr<api::SettingsController> settings_ctrl_;
};

} // namespace smartbotic::webserver