WEB Session Lifecycle Observability + Bridge Recovery Drafts

Co-Authored-By: brekotis <93345790+brekotis@users.noreply.github.com>

src/config/load/strict_keys.rs

@@ -344,6 +344,7 @@ const WEB_TIMEOUTS_CONFIG_KEYS: &[&str] = &[
     "long_poll_secs",
     "bridge_request_secs",
     "bridge_retry_secs",
+    "bridge_recovery_secs",
     "carrier_probe_coalesce_ms",
     "lane_open_wait_secs",
     "carrier_health_secs",

src/config/load/validate_web/timeouts.rs

@@ -42,6 +42,9 @@ pub(super) fn validate(timeouts: &WebTimeoutsConfig) -> Result<()> {
     if !(1..=300).contains(&timeouts.bridge_retry_secs) {
         return config_error("web.timeouts.bridge_retry_secs must be within [1, 300]");
     }
+    if !(1..=60).contains(&timeouts.bridge_recovery_secs) {
+        return config_error("web.timeouts.bridge_recovery_secs must be within [1, 60]");
+    }
     if timeouts.bridge_request_secs > timeouts.bridge_retry_secs {
         return config_error("web.timeouts.bridge_request_secs must not exceed bridge_retry_secs");
     }

src/config/tests/load_basic_tests/web_tests.rs

@@ -194,7 +194,7 @@ fn web_carriers_reject_true_empty_and_duplicates() {
 fn web_carrier_and_bridge_deadlines_are_configurable() {
     let configured = WEB_CONFIG.replace(
         "[[web.vhosts]]",
-        "[web.timeouts]\ncarrier_negotiation_deadlines_secs = [1, 2, 4, 9]\ncarrier_learning_secs = 30\nbridge_request_secs = 7\nbridge_retry_secs = 41\ncarrier_probe_coalesce_ms = 4\n\n[[web.vhosts]]",
+        "[web.timeouts]\ncarrier_negotiation_deadlines_secs = [1, 2, 4, 9]\ncarrier_learning_secs = 30\nbridge_request_secs = 7\nbridge_retry_secs = 41\nbridge_recovery_secs = 13\ncarrier_probe_coalesce_ms = 4\n\n[[web.vhosts]]",
     );
     let config = load_config_from_temp_toml(&configured);
     assert_eq!(
@@ -204,6 +204,7 @@ fn web_carrier_and_bridge_deadlines_are_configurable() {
     assert_eq!(config.web.timeouts.carrier_learning_secs, 30);
     assert_eq!(config.web.timeouts.bridge_request_secs, 7);
     assert_eq!(config.web.timeouts.bridge_retry_secs, 41);
+    assert_eq!(config.web.timeouts.bridge_recovery_secs, 13);
     assert_eq!(config.web.timeouts.carrier_probe_coalesce_ms, 4);
 }
 
@@ -211,13 +212,14 @@ fn web_carrier_and_bridge_deadlines_are_configurable() {
 fn web_bridge_deadlines_are_known_in_strict_mode() {
     let configured = WEB_CONFIG.replace(
         "[[web.vhosts]]",
-        "[web.timeouts]\nbridge_request_secs = 7\nbridge_retry_secs = 41\ncarrier_probe_coalesce_ms = 4\n\n[[web.vhosts]]",
+        "[web.timeouts]\nbridge_request_secs = 7\nbridge_retry_secs = 41\nbridge_recovery_secs = 13\ncarrier_probe_coalesce_ms = 4\n\n[[web.vhosts]]",
     );
     let configured = format!("[general]\nconfig_strict = true\n{configured}");
     let config = load_config_from_temp_toml(&configured);
 
     assert_eq!(config.web.timeouts.bridge_request_secs, 7);
     assert_eq!(config.web.timeouts.bridge_retry_secs, 41);
+    assert_eq!(config.web.timeouts.bridge_recovery_secs, 13);
     assert_eq!(config.web.timeouts.carrier_probe_coalesce_ms, 4);
 }
 
@@ -228,6 +230,8 @@ fn web_bridge_deadlines_are_bounded_and_ordered() {
         ("bridge_request_secs", "61"),
         ("bridge_retry_secs", "0"),
         ("bridge_retry_secs", "301"),
+        ("bridge_recovery_secs", "0"),
+        ("bridge_recovery_secs", "61"),
         ("carrier_probe_coalesce_ms", "11"),
     ] {
         let invalid = WEB_CONFIG.replace(

src/config/types/web.rs

@@ -301,6 +301,9 @@ pub struct WebTimeoutsConfig {
     /// Absolute generated-bridge budget for one retryable HTTP operation.
     #[serde(default = "default_web_bridge_retry_secs")]
     pub bridge_retry_secs: u64,
+    /// Absolute post-commit budget for one surviving bridge recovery epoch.
+    #[serde(default = "default_web_bridge_recovery_secs")]
+    pub bridge_recovery_secs: u64,
     /// Optional delay for coalescing the first OPEN with immediate DATA.
     #[serde(default = "default_web_carrier_probe_coalesce_ms")]
     pub carrier_probe_coalesce_ms: u64,
@@ -361,6 +364,7 @@ impl Default for WebTimeoutsConfig {
             long_poll_secs: default_web_long_poll_timeout_secs(),
             bridge_request_secs: default_web_bridge_request_secs(),
             bridge_retry_secs: default_web_bridge_retry_secs(),
+            bridge_recovery_secs: default_web_bridge_recovery_secs(),
             carrier_probe_coalesce_ms: default_web_carrier_probe_coalesce_ms(),
             lane_open_wait_secs: default_web_lane_open_wait_secs(),
             carrier_health_secs: default_web_carrier_health_secs(),

src/config/types/web/defaults.rs

@@ -88,6 +88,7 @@ u64_default!(default_web_stream_first_byte_secs, 30);
 u64_default!(default_web_long_poll_timeout_secs, 25);
 u64_default!(default_web_bridge_request_secs, 10);
 u64_default!(default_web_bridge_retry_secs, 90);
+u64_default!(default_web_bridge_recovery_secs, 15);
 u64_default!(default_web_carrier_probe_coalesce_ms, 0);
 u64_default!(default_web_lane_open_wait_secs, 2);
 u64_default!(default_web_carrier_health_secs, 30);