Remove the no-op MAINTENANCE_MODE setting (closes #317)

MAINTENANCE_MODE only made the healthcheck report
maintenanceMode: true; no request was ever served differently. It is
removed from the configuration, the healthcheck JSON, the README and
the Settings page, together with the unused Server.MaintenanceMode
method and the healthcheck's Config dependency, which existed only to
read it. A leftover MAINTENANCE_MODE in the environment is not read,
so it is ignored rather than refused.

Model: opus-5-5

internal/config/config.go

@@ -149,7 +149,6 @@ type ConfigParams struct {
 type Config struct {
 	DataDir         string
 	Debug           bool
-	MaintenanceMode bool
 	Environment     string
 	MetricsPassword string
 	MetricsUsername string
@@ -658,11 +657,6 @@ func loadFromEnv() (*Config, error) {
 		return nil, err
 	}
 
-	maintenanceMode, err := envBool("MAINTENANCE_MODE", false)
-	if err != nil {
-		return nil, err
-	}
-
 	retentionSweepInterval, err := envPositiveDuration(
 		"RETENTION_SWEEP_INTERVAL",
 		defaultRetentionSweepInterval,
@@ -712,7 +706,6 @@ func loadFromEnv() (*Config, error) {
 	return &Config{
 		DataDir:                DataDir(),
 		Debug:                  debug,
-		MaintenanceMode:        maintenanceMode,
 		Environment:            environment,
 		MetricsUsername:        metricsUsername,
 		MetricsPassword:        metricsPassword,
@@ -799,7 +792,6 @@ func New(lc fx.Lifecycle, params ConfigParams) (*Config, error) {
 		// host can reach the admin UI.
 		"bindAddress", s.BindAddress,
 		"debug", s.Debug,
-		"maintenanceMode", s.MaintenanceMode,
 		"dataDir", s.DataDir,
 		"retentionSweepInterval", s.RetentionSweepInterval.String(),
 		// Logged because a perfectly valid non-positive value here