bin/background-worker: Remove panic behavior

The only cases of our code hitting the `panic!()` in production so far have been because of a read-only Postgres database during database server maintenance. The r2d2 connection pool automatically recovers from this though.

In other words: there is no benefit in us panicking after five failures if in all observed instances it would have been sufficient to wait for the database to recover.

Since the startup on production takes multiple minutes due to the index repository cloning it seems somewhat bad to restart the whole dyno after just five seconds of database connection issues.

Author: Turbo87

src/bin/background-worker.rs

@@ -99,16 +99,9 @@ fn main() -> anyhow::Result<()> {
 
     info!("Runner booted, running jobs");
 
-    let mut failure_count = 0;
-
     loop {
-        if let Err(e) = runner.run_all_pending_jobs() {
-            failure_count += 1;
-            if failure_count < 5 {
-                warn!(?failure_count, err = ?e, "Error running jobs -- retrying");
-            } else {
-                panic!("Failed to begin running jobs 5 times. Restarting the process");
-            }
+        if let Err(err) = runner.run_all_pending_jobs() {
+            warn!(%err, "Failed to run background jobs");
         }
         sleep(Duration::from_secs(1));
     }