NuoDB Ops System Overview Graphs

This series of graphs are a collection of duplicate graphs for overview purposes.

Memory usage

This graph is a duplicate of the Memory usage graph.

Memory Usage.
Figure 1. Memory usage

Use Case

Memory used will tend to the set memory value for the --mem option. There will be a saw-tooth pattern where garbage collection will run. Flat lines must be investigated.

Graph Creation Detail

To create a graph using other tools, use the following detail:

  • Memory Usage: for the raw of HeapAllocated, average

CPU utilization per engine

This graph plots the total CPU used per Storage Manager (SM) and Transaction Engine (TE) normalized to max 100%.

CPU utilization per engine.
Figure 2. CPU utilization per engine

Use Case

The CPU usage per engine should not exceed 80% of max CPU. Anything over 80% must be investigated.

Graph Creation Detail

To create a graph using other tools, use the following detail:

  • PercentCpuTime: for the norm of PercentCpuTime, average

Transaction rate

This graph plots the rate of transactions committed per second. This is similar to Commits in SQL group of graphs.

Transaction rate.
Figure 3. Transaction rate

Use Case

Typically, the transaction rate depends on the workload. A heavier workload will mean a higher transaction rate. Longer periods where individual TEs have a high transaction rate may need investigation.

Graph Creation Detail

To create a graph using other tools, use the following detail:

  • Transaction rate: for the rate of Commits for TEs, average

Aggregate transaction rate

The sum of all the transactions per second across all TEs and the total number of TEs.

Aggregate Transaction rate.
Figure 4. Aggregate Transaction rate

Use Case

Typically, the aggregate transaction rate depends on the workload. A heavier workload will mean a higher aggregate transaction rate. Longer periods where there is a high aggregate transaction rate may need investigation.

Graph Creation Details

To create a graph using other tools, use the following details:

  • Aggregate Transaction Rate: sum the rate of Commits from each TE

  • Number of TEs: sum the value of Milliseconds from each TE

Client connections per engine

The number of client connections per TE. Similar to Connections in SQL graphs.

Client Connections per Engine.
Figure 5. Client Connections per Engine

Use case

Typically, client connections per SM and TE depend on application configuration and should be evenly spread across all TEs (except for tie-breaker TEs).

Graph Creation Detail

To create a graph using other tools, use the following detail:

  • client connections: for the raw number of ClientCncts, average.