If your postfix is running but mail isn't moving, check the /opt/zimbra/data/postfix/spool directory. When a ₹50,000 ($542) marketing blast hits a poorly tuned MTA, the deferred queue bloats until IOPS starvation kills the disk performance. The server isn't "broken"—it’s choking on its own data. If you don't manually purge the "maildrop" or increase the message_size_limit, your ₹1.2 Cr ($130,222) enterprise flow stays at a dead stop. This isn't a software bug; it's a configuration "Heart Attack."

It’s the "Ghost Service" syndrome. The mailboxd process might be "running" but the Java Virtual Machine (JVM) is stuck in a Full GC (Garbage Collection) loop. It’s unresponsive but the PID still exists, so Zimbra thinks it’s fine. Run jstat -gc right now. If your heap is at 99%, your server is a vegetable. We see this happen when admins forget to tune mailboxd_java_heap_memory_percent. It’s a ₹0 fix that prevents a ₹5 Lakh ($5,425) total system freeze.

If your admin console is slow and the stats aren't updating, your zmlogger database is likely borked. This usually happens after a hard reboot or a power failure in the datacenter. The syslog is trying to write to a corrupted MariaDB table, and the resulting I/O Wait drags down the entire CPU. Most people ignore it until the disk fills up with error logs. You have to run zmdbintegrityreport or drop the logger database entirely. It’s a messy, ₹0 "Manual Surgery" that saves you from a ₹1.5 Lakh ($1,627) "Emergency Support" call later.

Because you’re likely running ClamAV on a server with less than 16GB of RAM. In 2026, virus definitions are too massive for "budget" VPS builds. Amavis tries to scan a 5MB attachment, runs out of memory, swaps to disk, and effectively kills the server’s "Responsiveness." If you don't have the RAM Overhead, you shouldn't be running on-box scanning. Move the heavy lifting to a Spam/Virus Gateway. It’s a ₹25,000 ($271) hardware upgrade that prevents a systemic collapse of your ₹10 Lakh ($10,851) sales team.

If you just renewed your SSL but mobile users are still seeing "Security Error," you forgot the Root/Intermediate CA bundle. Zimbra’s zmcertmgr is notoriously picky. If the chain isn't perfect, the Nginx Proxy will serve a broken certificate, and every IMAP/POP3 client will disconnect. It’s a "Silent Failure"—the server looks fine, but the "Gate" is locked. We spend half our time fixing these "Chain Mismatches" for clients who thought they could "DIY" their security for ₹0.

If you’re running Zimbra Network Edition and your /opt/zimbra/redolog is 500GB+, your backup process is failing. Redo logs are supposed to roll over, but if the Metadata Sync hangs, they just keep growing until they consume every spare block on the SAN. This is "Storage Suicide." If you hit 0 bytes free, the database won't just stop—it might corrupt. This is the difference between a ₹3,500 ($38) storage bill and a ₹12 Lakh ($13,022) data recovery disaster.