ECS 2.1 – Monitor resources: hardware, network traffic, disk bandwidth, node and process health
Table of Contents
Resource monitoring includes:
Each of these monitoring areas has a corresponding page under the ECS Portal Monitor menu.
See the following articles for information about other ECS Portal monitoring features:
- Monitor storage: metering and capacity
- Monitor services: chunks, erasure coding, geo-replication, and recovery status
- Monitor events: audit portal, API, and CLI events and system alerts
The ECS Portal monitoring pages share a set of common interactions. These are:
- Search: the search icon appears when you can narrow monitoring results by matching search text to result rows containing the search text.
- Refresh: the refresh icon allows you to update the monitoring display with the latest data.
- Filter: fill in filter fields and the date range and select Filter to display result rows that match all filter fields. The default date range is always yesterday and today.
- Sort by column: Select a column head once to sort the results by that column in ascending order. Select again to change the sort order to descending order.
- Drill down displays with breadcrumbs: Breadcrumbs let you quickly drill up when you have drilled down into detail screens. See the example below.
- History charts with left to right mouse-overs: Get detailed charts showing hourly snapshots for the last five days worth of data which you can browse through using your mouse as a left-to-right chart cursor. See the example below.
Highlighted text in a table row indicates a link to a detail display. Selecting the link drills down to the next level of detail. On drill down displays, a path string shows your current location in the sequence of drill down displays. This path string is called a breadcrumb trail or breadcrumbs for short. Selecting any highlighted breadcrumb jumps up to the associated display.
When you select a History button, all available charts for that row display below the table. Mouse over a chart from left to right to see a vertical line that helps you find a specific date-time point on the chart. A pop-up display shows the value and timestamp for that point.Back to Top
Hardware health is designated by three states:
- Good: The hardware component is in normal operating condition.
- Suspect: Either the hardware component is transitioning from good to bad because of decreasing hardware metrics, or there is a problem with a lower-level hardware component, or the hardware is not detectable by the system because of connectivity problems.
- Bad: The hardware needs replacement.
In the case of disks, these states have the following meanings:
- Good: The system is actively reading from and writing to the disk.
- Suspect: The system no longer writes to the disk but will read from it. Note that "swarms" of suspect disks are likely caused by connectivity problems at a node. These disks will transition back to Good when the connectivity issues clear up.
- Bad: The system neither reads from nor writes to the disk. Replace the disk. Once a disk has been identified as bad by the ECS system, it cannot be reused anywhere in the ECS system. Because of ECS data protection, when a disk fails, copies of the data that was once on the disk are recreated on other disks in the system. A bad disk only represents a loss of capacity to the system--not a loss of data. When the disk is replaced, the new disk does not have data restored to it. It simply becomes raw capacity for the system.
- Select .
- Locate the table row for the target storage pool.
- Optionally, select a storage pool name to drill down to the node display.
- Optionally, select a node endpoint to drill down to the disk display.
See Replace an ECS storage disk in a U-Series Appliancefor information on using disk IDs to help identify disks needing replacement.
Thepage provides network traffic metrics at the virtual data center or the individual node level. The charts show data for the last seven days.
- Select .
- Locate the target VDC name.
- Optionally, select the VDC name to drill down to the nodes display.
History button ofr the target VDC or node.
Thepage provides disk use metrics at the virtual data center or the individual node level. There is one row for read and another for write for each VDC or node. The charts show data for the last seven days.
- Select .
- Locate the target VDC name and either the Read or Write table row for that VDC.
- Optionally, select the Node Count to drill down to a table with rows for the nodes in the VDC.
- Select the
History button for the VDC or node.
Thepage provides metrics that can help assess the health of the VDC, node, or node process.
- Locate the table row for the target VDC.
- Optionally, select the VDC name to drill down to a table with rows for each node in the VDC.
- Optionally, select the a node endpoint to drill down to a table with rows for each process running on the node.
- Select the
History button for the target VDC, node, or process.