Health

HPE Morpheus Software Health

../../_images/morpheusHealth500.png

The HPE Morpheus Software Health section provides an overview of the health of your HPE Morpheus Software appliance. It includes an appliance health summary in the following areas:

  • CPU: Appliance CPU usage is checked. If usage is greater than 50%, this indicator will be in a yellow or warning state. If HPE Morpheus Software is unable to complete the check, it will be in a red or error state. Depending on appliance performance and how frequently this indicator is in a warning state, it may be necessary to upgrade to increase CPU. The Overall health indicator will mirror the CPU health indicator

  • Memory: If swap usage is above 60% or HPE Morpheus Software memory usage is above 95%, this indicator will be in a yellow or warning state. If HPE Morpheus Software is unable to complete the check for any reason, it will be in a red or error state. Depending on appliance performance and how frequently this indicator is in a warning state, it may be necessary to increase swap, upgrade the appliance to add memory, or consider a different appliance architecture for those using single-node appliances

  • Storage: If utilization of the filesystem mounted at “/” exceeds 80%, this indicator will be in a yellow warned status. Above 90% will put this indicator in red or error status

  • Database: The database is checked. If the number of database connections exceeds the configured maximum number of connections or if any test queries are reported as being slow, this indicator will be in a yellow or warning state. If HPE Morpheus Software is unable to communicate with the database, it will be in a red or error state. In the database section further down the page, you can check the number of maximum used connections against the number of max connections. In the case of database connections exceeding the maximum, consider increasing the maximum settings connection

  • Elastic: Elasticsearch is polled for the health status of each index. If any indices are not reporting a “green” health status, this indicator will be in a yellow or warning state.

  • Queues: RabbitMQ queues are checked. Any queues containing more than 1000 messages are considered to be in an error state. Appliance Queue health is given in a yellow or warning status when any queues are in such an error state. In the Queues section further down the page you can see the individual Queues listed and which have messages piling up. When the appliance is unable to complete the check for any reason, this indicator will be in a red or error state

Health Levels

Health levels provide a live representation of the current memory and CPU load on the appliance. In an HA appliance, this data is specific to the application node serving the request. The base product does not identify that node on this page. Use load-balancer access logs and appliance logs to correlate a request with an application node; do not depend on an unmaintained forum plugin for this operational decision.

  • Morpheus CPU: Instantaneous amount of CPU capacity in use by HPE Morpheus Software processes

  • System CPU: Instantaneous amount of CPU capacity in use by all processes

  • Morpheus Memory: JVM maximum heap for the HPE Morpheus Software application process on the application node serving the page (see the metric definitions below)

  • System Memory: Instantaneous amount of total system memory currently claimed (this is commonly a high percentage, see the TIP box below)

  • Used Swap: Instantaneous amount of total available system swap in use

  • Storage: The instantaneous percentage utilization of the filesystem mounted at “/”

Tip

High system-memory utilization alone does not establish appliance pressure because operating-system usage includes caches and processes outside the application JVM. Use the metric definitions and warning guidance below. If warnings persist with degraded behavior, collect a support bundle and open a case through the HPE Support Center.

Additional System Health Indices

Memory metric definitions

Displayed metric

Current calculation

Scope and interpretation

Morpheus Memory

JVM maximum memory (Runtime.maxMemory)

Upper heap limit for the HPE Morpheus Software application process on the application node serving the page.

Morpheus Used Memory

JVM total memory minus JVM free memory

Heap currently committed to the JVM and occupied. This is not total resident memory for every appliance service.

Morpheus Free Memory

Runtime.freeMemory

Free space inside the currently committed JVM heap, not all memory available before the heap reaches its maximum.

Morpheus Memory Usage

Morpheus Used Memory / JVM total memory × 100

Percentage of the currently committed heap, not of Morpheus Memory (the maximum heap).

System Memory

Operating-system total physical memory

Host or container-visible physical-memory scope for the current application node.

System Used Memory

System Memory minus System Free Memory

Includes the OS, caches, and all visible processes.

System Free Memory

Operating-system free physical memory

Does not represent JVM free heap.

System Memory Usage

(System Memory - System Free Memory) / System Memory × 100

Whole node/container-visible usage.

System Swap / Free Swap

OS total swap / OS free swap

Swap visible to the current node.

The values do not form one arithmetic breakdown: JVM totalMemory is committed heap, JVM maxMemory is a limit, and system values include other processes and caches. They are sampled when health data is loaded. In HA deployments, each application node has its own JVM and system sample; compare the node that raised the warning rather than adding values across nodes.

Memory warning decisions

The current health implementation reports a warning when used swap is greater than 60% of total swap. If that test is not met, it reports a warning when Morpheus Used Memory is greater than 95% of the JVM’s currently committed total memory. These tests are ordered; the displayed message identifies the first matched condition. A collection error produces an error state rather than a utilization warning.

Treat an isolated high value as a point-in-time signal. Check whether the warning persists, review application performance and garbage-collection behavior, and collect a support bundle before changing memory limits. For sustained swap pressure, investigate host/container memory pressure and competing processes. For sustained Morpheus heap pressure with degraded behavior, engage support to review sizing before increasing the application memory limit. Health thresholds are appliance alarms and are unrelated to workload resize Guidance thresholds.

CPU

  • Processor Count

  • Process Time

  • Morpheus CPU

  • System CPU

  • System Load

MEMORY

  • Morpheus Memory

  • Morpheus Used Memory

  • Morpheus Free Memory

  • Morpheus Memory Usage

  • System Memory

  • System Used Memory

  • System Free Memory

  • System Memory Usage

  • System Swap

  • Free Swap

DATABASE

  • Lifetime Connections

  • Aborted Connections

  • Max Used Connections

  • Max Connections

  • Threads Running

  • Threads Connected

  • Slow Queries

  • Temp Tables

  • Key Reads

  • Handler Reads

  • Buffer Pool Free

  • Open Tables

  • Table Scans

  • Full Joins

  • Key Read Requests

  • Key Reads

  • Engine Waits

  • Lock Waits

  • Handler Reads

  • Engine IO Writes

  • Engine IO Reads

  • Engine IO Double Writes

  • Engine Log Writes

  • Engine Memory

  • Dictionary Memory

  • Buffer Pool Size

  • Free Buffers

  • Database Pages

  • Old Pages

  • Dirty Page Percent

  • Max Dirty Pages

  • Pending Reads

  • Insert Rate

  • Update Rate

  • Delete Rate

  • Read Rate

  • Buffer Hit Rate

  • Read Write Ratio

  • Uptime

ELASTIC

  • Status

  • Cluster

  • Node Count

  • Data Nodes

  • Shards

  • Primary Shards

  • Relocating Shards

  • Initializing

  • Unassigned

  • Pending Tasks

  • Active Shards

Note

Warning status is typical for Elasticsearch

Elastic Nodes

  • Node

  • Master

  • Location

  • Heap Usage

  • Memory Usage

  • CPU Usage

  • 1M Load

  • 5M Load

  • 15M Load

Elastic Indices

  • Health

  • Index

  • Status

  • Primary

  • Replicas

  • Doc

  • Count

  • Primary

  • Size

  • Total Size

Queues

  • Queue Count

  • Busy Queues

  • Error Queues

HPE Morpheus Software Logs

The HPE Morpheus Software logs section aggregates appliance-specific logs into one list. If needed, users can export the logs by clicking EXPORT. This action triggers a download containing the last 10,000 log entries as a .log file.

../../_images/healthlogs.png