OCI Load Balancer Backend Sets Performance Test

A backend set is a logical grouping of backend servers that handle the traffic routed through the load balancer. This also includes configuration settings that govern how that traffic is distributed and monitored. When you create a load balancer, one or more backend sets are defined, and then those backend sets are associated with listeners. A listener listens for incoming traffic on a specified port and protocol, and forwards it to a backend set for processing.

This test continuously monitors the backend sets in the target Oracle Cloud Infrastructure and for each backend set, this test reports the health, connection activity, data flow, HTTP response codes, and response times. This offers the administrators essential insights for maintaining availability, efficiency, and responsiveness.

Target of the test: Oracle Cloud Infrastructure Load Balancer

Agent deploying the test: A Remote Agent

Outputs of the test: One set of results for each Load Balancer:Backend set:Regionthat is being monitored.

Configurable parameters for the test

Parameter

Description

Test period

How often should the test be executed.

Host

The IP address of the target host to be monitored.

Tenancy ID

To monitor the target OCI Load Balancer, the eG agent has to be configured with the Tenancy ID of the user who is authorized to access the OCI Load Balancer. To know how to determine a Tenancy ID, refer to Generating an API Key.

Confirm Tenancy ID

Confirm the Tenancy ID by re-typing in the Confirm Tenancy ID text box.

Compartment ID

Specify the ID of the compartment in which the load balancer that you want to monitor is grouped, in the Compartment ID text box. To know how to view the Compartment ID, refer to Viewing the Compartment ID.

Confirm Compartment ID

Confirm the Compartment ID by re-typing in the Confirm Compartment ID text box.

User ID

Specify the ID of the user who is authorized to access the OCI Load Balancer in the User ID text box. You can either create a new user or use an existing user for this purpose. To create a new user and obtain the User ID of the user, refer to Pre-requisites for Monitoring Oracle Cloud Infrastructure Load Balancer.

Confirm User ID

Confirm the User ID by re-typing in the Confirm User ID text box.

Fingerprint

The fingerprint authenticates the API call, confirming its origin from the designated user i.e., the user who is authorized to access the target OCI Load Balancer. By default, the fingerprint is exposed while creating a new user and generating an API Key. Specify the fingerprint that you had noted down earlier while generating the API Key in the Fingerprint text box. To know how to identify the fingerprint, refer to Generating an API Key.

Confirm Fingerprint

Confirm the Fingerprint by re-typing in the Confirm Fingerprint text box.

Confirm Compartment ID

Confirm the Compartment ID by re-typing in the Confirm Compartment ID text box.

Private Key File Path

Specify the full path to the Private Key that you have downloaded while generating an API key in the Private Key File Path text box. By default, the Private key is downloaded as a .pem file. For example, your location can be: /opt/egurkha/oci/oci_key.pem

Regions

Specify the region where the load balancer and its associated resources are tied to in the Regions text box. To know the region, refer to Generating an API Key.

Measurements made by the test

Measurement

Description

Measurement Unit

Interpretation

Active connections

Indicates the number of connections that were open to this region.

Number

 

Closed connections

Indicates the number of connections that were closed to this region.

Number

The value of this measure includes the connections that were closed due to errors because the connections were no longer required.

Backend servers

Indicates the total number of backend servers connected to this region.

Number

 

Timeouts across all backend servers

Indicates the total number of backend servers that had timed out in this region.

Number

 

Data received rate

Indicates the rate at which data was received by this region.

KB/sec

Compare the value of this measure across the backend sets to know which backend set is receiving maximum data.

Data sent rate

Indicates the rate at which data was transmitted from this region.

KB/sec

Compare the value of this measure across the backend sets to know which backend set is transmitting maximum data.

HTTP requests

Indicates the number of HTTP requests processed by this region.

Number

 

Total responses

Indicates the total number of responses made by this region.

Number

 

2xx SLA achievement

Indicates the percentage of of successful HTTP responses (status codes in the 2xx range) delivered by this region.

Percent

 

HTTP 200 responses

Indicates the total number of HTTP 200 responses made by this backend set.

Number

 

HTTP 2xx responses

Indicates the total number of 2xx responses made by this region.

Number

The value of this measure denotes the number of responses by this backend set that were successful.

HTTP 3xx responses

Indicates the total number of 3xx responses made by this region.

Number

The value of this measure denotes the number of responses by this backend set that needs further action.

HTTP 4xx responses

Indicates the total number of 4xx responses made by this region.

Number

The value of this measure denotes the number of responses by this backend set that has bad syntax or cannot be fulfilled.

HTTP 5xx responses

Indicates the total number of 5xx responses made by this region.

Number

The value of this measure denotes the number of responses by this backend set where the server failed to fulfill a valid request.

HTTP 502 responses

Indicates the number of HTTP 502 responses received by this region.

Number

When this response code is encountered by the backend set, the load balancer is usually healthy, but one or more of the backend servers are failing or misbehaving.

HTTP 504 responses

Indicates the number of HTTP 504 responses received by this region.

Number

When this response code is encountered by the backend set, it means the load balancer gave up waiting for a response from the backend before the timeout threshold.

Invalid header responses

Indicates the number of responses in this region with invalid headers.

Number

This means the load balancer detected responses from backend servers that did not conform to HTTP header rules.

A very low value of this measure is desired.

Keep-alive connections

Indicates the total number of keep-alive connections in this region.

Number

Keep-Alive connections allow a single TCP connection to stay open for multiple HTTP requests/responses, reducing the overhead of repeatedly opening and closing new connections. This helps in reducing latency.

Average HTTP response time

Indicates the average time taken to respond to a HTTP request in this region.

Milliseconds

 

Average TCP response time

Indicates the average time taken to respond to a TCP request in this region.

Milliseconds

 

Unhealthy backend servers

Indicates the number of unhealthy backend servers in this region.

Number