fix: silent NWC auto-payment failure messaging in invoice.created fallback

Co-authored-by: userAdityaa <aditya.chaudhary1558@gmail.com>
Co-committed-by: userAdityaa <aditya.chaudhary1558@gmail.com>

backend/spec/infra.md

@@ -15,12 +15,15 @@ Members:
 ## `pub async fn start(self)`
 
 - Subscribes to `command.notify`
+- On startup, schedules delayed sync retries for relays whose `sync_error` is non-empty.
 - Loops on `rx.recv()`, calling `handle_activity` for each received `Activity`.
 
 ## `async fn handle_activity(&self, activity: &Activity)`
 
-- For `create_relay`, `update_relay`, `activate_relay`, or `deactivate_relay` activity, calls `sync_and_report`.
-- All other activity types are ignored (e.g. `fail_relay_sync`, `complete_relay_sync`).
+- For `create_relay`, `update_relay`, `activate_relay`, or `deactivate_relay` activity, calls `sync_and_report` immediately.
+- For `fail_relay_sync`, schedules a delayed retry using exponential backoff based on consecutive failures for the relay.
+- Retry scheduling stops after the configured max attempts to avoid infinite retry loops.
+- Other activity types are ignored (e.g. `complete_relay_sync`).
 
 ## `async fn sync_and_report(&self, relay: &Relay, is_new: bool)`
 

backend/src/billing.rs

@@ -17,6 +17,9 @@ type HmacSha256 = Hmac<Sha256>;
 const STRIPE_API: &str = "https://api.stripe.com/v1";
 const COINBASE_SPOT_API: &str = "https://api.coinbase.com/v2/prices";
 const WEBHOOK_TOLERANCE_SECS: i64 = 300;
+const MANUAL_LIGHTNING_PAYMENT_DM: &str = "Payment is due for your relay subscription. Please visit the application to complete a manual Lightning payment.";
+const NWC_ERROR_DM_PREFIX: &str = "NWC auto-payment failed:";
+const NWC_ERROR_DM_MAX_CHARS: usize = 240;
 
 #[derive(Debug)]
 pub enum InvoiceLookupError {
@@ -80,7 +83,6 @@ pub struct Billing {
     nwc_url: String,
     stripe_secret_key: String,
     stripe_webhook_secret: String,
-    btc_quote_api_base: String,
     http: reqwest::Client,
     query: Query,
     command: Command,
@@ -98,13 +100,10 @@ impl Billing {
         if stripe_webhook_secret.trim().is_empty() {
             panic!("missing STRIPE_WEBHOOK_SECRET environment variable");
         }
-        let btc_quote_api_base =
-            std::env::var("BTC_PRICE_API_BASE").unwrap_or_else(|_| COINBASE_SPOT_API.to_string());
         Self {
             nwc_url,
             stripe_secret_key,
             stripe_webhook_secret,
-            btc_quote_api_base,
             http: reqwest::Client::new(),
             query,
             command,
@@ -360,6 +359,8 @@ impl Billing {
             return Ok(());
         };
 
+        let mut nwc_error_for_dm: Option<String> = None;
+
         // 1. NWC auto-pay: if the tenant has a nwc_url
         if !tenant.nwc_url.is_empty() {
             match self
@@ -376,6 +377,14 @@ impl Billing {
                     self.command
                         .set_tenant_nwc_error(&tenant.pubkey, &error_msg)
                         .await?;
+                    tracing::warn!(
+                        error = %e,
+                        tenant_pubkey = %tenant.pubkey,
+                        stripe_customer_id,
+                        invoice_id,
+                        "nwc auto-payment failed for invoice.created"
+                    );
+                    nwc_error_for_dm = summarize_nwc_error_for_dm(&error_msg);
                     // Fall through to next option
                 }
             }
@@ -390,12 +399,8 @@ impl Billing {
         }
 
         // 3. Manual payment: send a DM
-        self.robot
-            .send_dm(
-                &tenant.pubkey,
-                "Payment is due for your relay subscription. Please visit the application to complete a manual Lightning payment.",
-            )
-            .await?;
+        let dm_message = manual_lightning_payment_dm(nwc_error_for_dm.as_deref());
+        self.robot.send_dm(&tenant.pubkey, &dm_message).await?;
 
         Ok(())
     }
@@ -1057,7 +1062,7 @@ impl Billing {
     }
 
     async fn fetch_btc_spot_price(&self, currency: &str) -> Result<f64> {
-        fetch_btc_spot_price_from_base(&self.http, &self.btc_quote_api_base, currency).await
+        fetch_btc_spot_price_from_base(&self.http, COINBASE_SPOT_API, currency).await
     }
 
     fn currency_minor_exponent(currency: &str) -> Result<u8> {
@@ -1101,6 +1106,31 @@ pub async fn fetch_btc_spot_price_from_base(
     Ok(amount)
 }
 
+fn summarize_nwc_error_for_dm(error: &str) -> Option<String> {
+    let normalized = error.split_whitespace().collect::<Vec<_>>().join(" ");
+    if normalized.is_empty() {
+        return None;
+    }
+
+    if normalized.chars().count() <= NWC_ERROR_DM_MAX_CHARS {
+        return Some(normalized);
+    }
+
+    let prefix_len = NWC_ERROR_DM_MAX_CHARS.saturating_sub(3);
+    let mut truncated = normalized.chars().take(prefix_len).collect::<String>();
+    truncated.push_str("...");
+    Some(truncated)
+}
+
+fn manual_lightning_payment_dm(nwc_error: Option<&str>) -> String {
+    match nwc_error {
+        Some(error) if !error.is_empty() => {
+            format!("{MANUAL_LIGHTNING_PAYMENT_DM}\n\n{NWC_ERROR_DM_PREFIX} {error}")
+        }
+        _ => MANUAL_LIGHTNING_PAYMENT_DM.to_string(),
+    }
+}
+
 pub fn fiat_minor_to_msats_from_quote(
     amount_due_minor: i64,
     currency: &str,

backend/src/infra.rs

@@ -1,10 +1,15 @@
 use anyhow::Result;
 use nostr_sdk::prelude::*;
+use std::time::Duration;
 
 use crate::command::Command;
 use crate::models::{Activity, RELAY_STATUS_DELINQUENT, RELAY_STATUS_INACTIVE, Relay};
 use crate::query::Query;
 
+const RELAY_SYNC_RETRY_BASE_DELAY_SECS: u64 = 30;
+const RELAY_SYNC_RETRY_MAX_DELAY_SECS: u64 = 15 * 60;
+const RELAY_SYNC_RETRY_MAX_ATTEMPTS: usize = 6;
+
 #[derive(Clone)]
 pub struct Infra {
     api_url: String,
@@ -48,6 +53,10 @@ impl Infra {
     pub async fn start(self) {
         let mut rx = self.command.notify.subscribe();
 
+        if let Err(e) = self.schedule_startup_retries().await {
+            tracing::error!(error = %e, "failed to schedule relay sync retries on startup");
+        }
+
         loop {
             match rx.recv().await {
                 Ok(activity) => {
@@ -66,15 +75,84 @@ impl Infra {
     async fn handle_activity(&self, activity: &Activity) -> Result<()> {
         let needs_sync = should_sync_relay_activity(activity.activity_type.as_str());
 
-        if needs_sync {
-            let Some(relay) = self.query.get_relay(&activity.resource_id).await? else {
-                return Ok(());
-            };
-
-            let is_new = relay.synced == 0;
-            self.sync_and_report(&relay, is_new).await;
+        if !needs_sync || activity.resource_type != "relay" {
+            return Ok(());
         }
 
+        if activity.activity_type == "fail_relay_sync" {
+            self.schedule_relay_sync_retry(&activity.resource_id, "activity")
+                .await?;
+            return Ok(());
+        }
+
+        let Some(relay) = self.query.get_relay(&activity.resource_id).await? else {
+            return Ok(());
+        };
+
+        let is_new = relay.synced == 0;
+        self.sync_and_report(&relay, is_new).await;
+
+        Ok(())
+    }
+
+    async fn schedule_startup_retries(&self) -> Result<()> {
+        let relays = self.query.list_relays_with_sync_error().await?;
+
+        for relay in relays {
+            self.schedule_relay_sync_retry(&relay.id, "startup").await?;
+        }
+
+        Ok(())
+    }
+
+    async fn schedule_relay_sync_retry(&self, relay_id: &str, source: &str) -> Result<()> {
+        let activities = self.query.list_activity_for_relay(relay_id).await?;
+        let consecutive_failures = consecutive_sync_failures(&activities);
+
+        let Some(delay) = relay_sync_retry_delay(consecutive_failures) else {
+            tracing::warn!(
+                relay = relay_id,
+                consecutive_failures,
+                max_attempts = RELAY_SYNC_RETRY_MAX_ATTEMPTS,
+                "relay sync retries exhausted; awaiting manual intervention"
+            );
+            return Ok(());
+        };
+
+        tracing::info!(
+            relay = relay_id,
+            source,
+            consecutive_failures,
+            delay_secs = delay.as_secs(),
+            "scheduled relay sync retry"
+        );
+
+        let relay_id = relay_id.to_string();
+        let infra = self.clone();
+
+        tokio::spawn(async move {
+            tokio::time::sleep(delay).await;
+
+            if let Err(e) = infra.retry_relay_sync(&relay_id).await {
+                tracing::error!(relay = %relay_id, error = %e, "relay sync retry task failed");
+            }
+        });
+
+        Ok(())
+    }
+
+    async fn retry_relay_sync(&self, relay_id: &str) -> Result<()> {
+        let Some(relay) = self.query.get_relay(relay_id).await? else {
+            return Ok(());
+        };
+
+        if relay.sync_error.trim().is_empty() {
+            tracing::debug!(relay = %relay.id, "skip relay sync retry; relay has no sync_error");
+            return Ok(());
+        }
+
+        let is_new = relay.synced == 0;
+        self.sync_and_report(&relay, is_new).await;
         Ok(())
     }
 
@@ -258,6 +336,28 @@ fn relay_sync_body(
 fn should_sync_relay_activity(activity_type: &str) -> bool {
     matches!(
         activity_type,
-        "create_relay" | "update_relay" | "activate_relay" | "deactivate_relay"
+        "create_relay" | "update_relay" | "activate_relay" | "deactivate_relay" | "fail_relay_sync"
     )
 }
+
+fn consecutive_sync_failures(activities: &[Activity]) -> usize {
+    activities
+        .iter()
+        .take_while(|activity| activity.activity_type == "fail_relay_sync")
+        .count()
+}
+
+fn relay_sync_retry_delay(consecutive_failures: usize) -> Option<Duration> {
+    let retry_attempt = consecutive_failures.max(1);
+    if retry_attempt > RELAY_SYNC_RETRY_MAX_ATTEMPTS {
+        return None;
+    }
+
+    let exponent = (retry_attempt - 1).min(31);
+    let multiplier = 1u64 << exponent;
+    let delay_secs = RELAY_SYNC_RETRY_BASE_DELAY_SECS
+        .saturating_mul(multiplier)
+        .min(RELAY_SYNC_RETRY_MAX_DELAY_SECS);
+
+    Some(Duration::from_secs(delay_secs))
+}

backend/src/query.rs

@@ -94,6 +94,23 @@ impl Query {
         Ok(rows)
     }
 
+    pub async fn list_relays_with_sync_error(&self) -> Result<Vec<Relay>> {
+        let rows = sqlx::query_as::<_, Relay>(
+            "SELECT id, tenant, schema, subdomain, plan, stripe_subscription_item_id,
+                     status, sync_error,
+                     info_name, info_icon, info_description,
+                     policy_public_join, policy_strip_signatures,
+                     groups_enabled, management_enabled, blossom_enabled,
+                     livekit_enabled, push_enabled, synced
+              FROM relay
+              WHERE TRIM(sync_error) != ''
+              ORDER BY id",
+        )
+        .fetch_all(&self.pool)
+        .await?;
+        Ok(rows)
+    }
+
     pub async fn list_relays_for_tenant(&self, tenant_id: &str) -> Result<Vec<Relay>> {
         let rows = sqlx::query_as::<_, Relay>(
             "SELECT id, tenant, schema, subdomain, plan, stripe_subscription_item_id,