From 81527ca611f4da931e4dc0387405667110f9d222 Mon Sep 17 00:00:00 2001
From: Pavan Nagaraj <pavan.nagaraj@spectrocloud.com>
Date: Wed, 1 Jul 2026 20:52:45 +0530
Subject: [PATCH] feat(support-bundle): collect MongoDB per-collection sizes in
 infra and edge bundles

Add per-database and per-collection storage/data sizes and document counts
(mongo/db-collection-sizes.txt) to both the infra and edge support bundles,
via listDatabases + db.stats()/collection.stats(). This surfaces WHICH
collections consume disk (events, audits, metrics, cluster summaries, etc.),
the recurring "what is eating MongoDB storage" triage question.

Follow-up to the per-pod df collection; motivated by SCS-4625 / SUS-1801.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
---
 support-bundle/README-edge.md          |  1 +
 support-bundle/README-infra.md         |  1 +
 support-bundle/support-bundle-edge.sh  | 11 +++++++++++
 support-bundle/support-bundle-infra.sh | 11 +++++++++++
 4 files changed, 24 insertions(+)

diff --git a/support-bundle/README-edge.md b/support-bundle/README-edge.md
index f1e4d9b..b2cde48 100644
--- a/support-bundle/README-edge.md
+++ b/support-bundle/README-edge.md
@@ -207,3 +207,4 @@ For Enterprise and PCG clusters, the script collects MongoDB replica set informa
 * `rs-conf.json`: Replica set configuration including member settings and priorities
 * `replication-info.txt`: Oplog information and replication window
 * `disk-usage.txt`: Actual `df -h /var/lib/mongodb` per mongo pod — the real on-disk size, which can differ from the requested PVC/PV size on thin-provisioned or fixed-disk-offering backends (e.g. CloudStack fixed offerings)
+* `db-collection-sizes.txt`: Per-database and per-collection storage/data sizes (MB) and document counts — shows which collections are consuming disk
diff --git a/support-bundle/README-infra.md b/support-bundle/README-infra.md
index 4941795..972219c 100644
--- a/support-bundle/README-infra.md
+++ b/support-bundle/README-infra.md
@@ -151,6 +151,7 @@ For Enterprise and PCG clusters, the script collects MongoDB replica set informa
 * `rs-conf.json`: Replica set configuration including member settings and priorities
 * `replication-info.txt`: Oplog information and replication window
 * `disk-usage.txt`: Actual `df -h /var/lib/mongodb` per mongo pod — the real on-disk size, which can differ from the requested PVC/PV size on thin-provisioned or fixed-disk-offering backends (e.g. CloudStack fixed offerings)
+* `db-collection-sizes.txt`: Per-database and per-collection storage/data sizes (MB) and document counts — shows which collections are consuming disk
 
 ## Important Notes
 
diff --git a/support-bundle/support-bundle-edge.sh b/support-bundle/support-bundle-edge.sh
index 730476b..104a0c2 100755
--- a/support-bundle/support-bundle-edge.sh
+++ b/support-bundle/support-bundle-edge.sh
@@ -682,6 +682,17 @@ function mongo-status() {
       echo
     } >> "${TMPDIR}/mongo/disk-usage.txt"
   done < <(printf '%s\n' "$MONGO_PODS")
+
+  techo "Collecting MongoDB database + collection sizes"
+  kubectl exec -n hubble-system "$MONGO_POD" -c mongo -- bash -c "$MONGO_CMD $MONGO_AUTH admin --quiet --eval '
+    db.adminCommand({listDatabases:1}).databases.forEach(function(d){
+      var s = db.getSiblingDB(d.name).stats(1024*1024);
+      print(\"DB \"+d.name+\" storageSize(MB)=\"+s.storageSize+\" dataSize(MB)=\"+s.dataSize);
+      db.getSiblingDB(d.name).getCollectionNames().forEach(function(c){
+        var cs = db.getSiblingDB(d.name).getCollection(c).stats(1024*1024);
+        print(\"  \"+d.name+\".\"+c+\" storageSize(MB)=\"+cs.storageSize+\" size(MB)=\"+cs.size+\" count=\"+cs.count);
+      });
+    });'" > "${TMPDIR}/mongo/db-collection-sizes.txt" 2>&1
 }
 
 function k8s-resources() {
diff --git a/support-bundle/support-bundle-infra.sh b/support-bundle/support-bundle-infra.sh
index e5d6e2e..92641c3 100755
--- a/support-bundle/support-bundle-infra.sh
+++ b/support-bundle/support-bundle-infra.sh
@@ -192,6 +192,17 @@ function mongo-status() {
       echo
     } >> "${TMPDIR}/mongo/disk-usage.txt"
   done < <(printf '%s\n' "$MONGO_PODS")
+
+  techo "Collecting MongoDB database + collection sizes"
+  kubectl exec -n hubble-system "$MONGO_POD" -c mongo -- bash -c "$MONGO_CMD $MONGO_AUTH admin --quiet --eval '
+    db.adminCommand({listDatabases:1}).databases.forEach(function(d){
+      var s = db.getSiblingDB(d.name).stats(1024*1024);
+      print(\"DB \"+d.name+\" storageSize(MB)=\"+s.storageSize+\" dataSize(MB)=\"+s.dataSize);
+      db.getSiblingDB(d.name).getCollectionNames().forEach(function(c){
+        var cs = db.getSiblingDB(d.name).getCollection(c).stats(1024*1024);
+        print(\"  \"+d.name+\".\"+c+\" storageSize(MB)=\"+cs.storageSize+\" size(MB)=\"+cs.size+\" count=\"+cs.count);
+      });
+    });'" > "${TMPDIR}/mongo/db-collection-sizes.txt" 2>&1
 }
 
 function k8s-resources() {