Monitor MongoDB: health and slow queries
The MongoDB menu monitors each client’s database: whether it is up, the slow queries in progress, replicas and cache. It shows up on clients with a connection string saved and is read-only on every plan: Kubepier writes and deletes nothing in the database.
What it does
| Plan | What it shows |
|---|---|
| Free | Whether the database is up or down, ping latency and version. |
| Pro and Team | Health view: ok, slow or down; connections; operations per second (two samples one second apart); replicas and replication lag; WiredTiger cache usage; current slow operations; the busiest collections (top) and the largest collections. |
When MongoDB shows slow or down
| State | When |
|---|---|
| Down | The ping gets no answer within 5 seconds |
| Slow | Any of: ping of 100 ms or more, an operation running 5 seconds or more, a replica 10 seconds or more behind the primary, or WiredTiger cache at 95% or more |
| Ok | None of the above |
The same thresholds apply on the web and on desktop, and the screen says which one was crossed.
Limits and timeouts
| Limit | Value |
|---|---|
| Slow operations listed | up to 30, running 1 second or more |
| Top collections | 10 |
| Largest collections | the 10 largest in the database saved on the client |
| Ping | no answer within 5 seconds = down |
What never shows in slow queries
In slow operations, the filter and command keep only fields and operators: every value becomes "?". Documents are never read.
Permissions you need on your side
| Panel | MongoDB user role |
|---|---|
| Status (Free) | Any user that authenticates (ping and buildInfo) |
| Server, replicas, slow operations and top | clusterMonitor on the admin database |
| Largest collections | read on the database saved on the client |
Each panel fails on its own when the role is missing and says which role it needs; the others keep working. On MongoDB Atlas, use a database user with clusterMonitor and read on the database.
What is not accepted
- Any write, delete, running a query or reading a document.
- The full view on Free.
Network and connection route
In the MongoDB form, pick the Connection route. Direct: allow the egress IPs (on Atlas, under Network Access → IP Access List). Through the cluster: for MongoDB on a private IP or private endpoint; replica sets and mongodb+srv:// work, resolved inside the cluster. On the Direct route, a host that resolves only to a private IP answers host_privado right away, without trying to connect. A connection closed by the server (conexao_encerrada) is usually a firewall without the egress IPs, the wrong TLS setting or port, or a connection limit: allow the IPs or switch to Through the cluster.
Common errors
- "Needs the clusterMonitor role": grant the role to the connection string user.
- Down: the server did not answer within 5 seconds: check the connection string and, on the web, the egress IPs in the Atlas Access List.