Monitoring Clusters and Nodes
You can use the Buckit Manager web interface to monitor the health and status of every cluster you manage, drill into individual nodes, and review per-node service logs. This page describes the monitoring views.
Clusters List
The Clusters page is the default landing page and lists every cluster registered with this manager. Each row summarizes one cluster:
Column |
Description |
|---|---|
Name |
The cluster’s display name. Clusters still running MinIO Community Edition are tagged with a MinIO pill. Select the name to open the cluster detail page. |
Pools |
The number of server pools in the cluster. |
Nodes (Online/Total) |
The number of nodes reporting online out of the total. The online count is highlighted when it is below the total. |
Drives (Ready/Total) |
The number of drives in the ready state out of the total, highlighted when below the total. |
Version |
The Buckit (or MinIO) version the cluster is running. |
Health |
An overall health pill: Healthy, Degraded, Critical, or Unknown. |
Used |
Used capacity out of usable capacity. |
Buckit Manager fetches fresh cluster state in the background. The page header shows how long ago the data was last fetched and provides a Refresh button to fetch on demand. If the data becomes stale, Buckit Manager refreshes it automatically.
From this page you can also start the Deploy new cluster and Import existing cluster flows. See Adding Clusters.
Cluster Details
Selecting a cluster opens the cluster detail page. When the page opens and at a regular interval while it is open, Buckit Manager refreshes the cluster state and re-probes connectivity so the view stays current.
The header shows the cluster name, an overall health pill, and the cluster metadata (version, node count, pool count, and erasure-coding parity). For clusters migrated from MinIO, the header also notes the original MinIO version.
Summary Cards
Three summary cards sit below the header:
Health — the overall health pill plus a breakdown of nodes online (and any degraded or offline) and drives ready (and any healing or failed).
Capacity — used capacity out of usable capacity, with a usage bar and percent used.
Pools — a per-pool health rollup. Each pool shows its own health pill and node and drive counts. When a cluster has many pools, the card shows the worst-health pools first and lets you expand to see the rest, so a problem is never hidden behind healthy pools.
Nodes Table
The nodes table lists every node in the cluster. Each row links to the node detail page and shows:
Column |
Description |
|---|---|
Host |
The node hostname. Select it to open the node detail page. |
Pool |
The server pool the node belongs to. |
State |
The node state: Online, Degraded, Offline, Unreachable, or Unknown. |
Version |
The Buckit version running on the node. |
Ping, SSH, S3 API, Console |
Connectivity probes. A green dot indicates the probe succeeded; a red mark indicates it failed. The probes cover ICMP/TCP reachability, a TCP connection to the SSH port, the S3 API, and the Buckit web console. |
Kernel |
The node’s kernel version. |
Uptime |
How long the node has been up. |
The table supports per-column filtering (host, pool, state, version, and kernel) and column sorting. By default, nodes are grouped by pool in ascending order, with hostnames sorted alphabetically within each pool.
Use the row checkboxes to select one or more hosts and run host-level operations from the action bar above the table. See Host Actions.
You can run cluster-wide operations from the Cluster Actions menu, and open cluster settings from the gear menu next to the cluster name. See Cluster and Node Operations. If the cluster is still running MinIO, the header also offers Migrate to Buckit and a link to open the MinIO console.
Node Details
Selecting a node opens the node detail page, which provides a focused view of a single host. Buckit Manager refreshes the node and re-runs its connectivity probes when the page opens and at a regular interval while it is open.
The page presents three cards:
System — operating system, kernel version and architecture, and the last boot time.
Hardware — CPU model, socket and core counts, clock speed, total and free RAM, and the primary network interface and link speed.
Connectivity — the same Ping, SSH, S3 API, and Console probe results shown in the nodes table, in an expanded form.
Note
Detailed hardware facts (CPU, memory, NIC) require valid admin credentials for the cluster. If admin credentials are not configured, Buckit Manager shows a note in the System card. See Connection and Application Settings.
Below the cards, the Drives table lists each data drive on the node with its mount point, device, size, used percentage, and status (Ready, Healing with percent complete, Failed, or Unknown). Boot drives are excluded.
The Actions menu on this page runs node-scoped operations. See Node Actions.
Service Logs
From the node Actions menu, select View service log to
open the per-node log viewer. Buckit Manager fetches recent journalctl
output for the Buckit service from the node over SSH.
The log viewer provides:
A Range selector (last 15 minutes, hour, 6 hours, 24 hours, or 7 days).
A Filter field for client-side substring matching.
A Refresh button to re-fetch the latest lines.
The log viewer performs a one-shot fetch each time you choose a range or select Refresh; it is not a continuous live tail. Lines that look like errors or warnings are color-coded.
Note
Viewing service logs requires SSH credentials for the cluster. If SSH is not configured, Buckit Manager prompts you to set it up. See Connection and Application Settings.
