This question can be interpreted two ways:

What to use for Health Checks (formally known as Golden Signals)?

The service can be configured with multiple health checks.  The results from all endpoints are checked during the automated testing.  Configure as many, or as few, signals as makes sense for this service.

Generally, we advise monitoring:

• Latency: what’s the response time of this service?  Visible results generally need to be under 200ms or the web experience starts to lag.

• Error Rate: are errors happening too frequently e.g. 5% of results are returning 5xx

• Traffic Levels: is the traffic volume within expected ranges for this time?

• Saturation: overall consumption of the service i.e. it’s at 25/50/75% of overall capacity

How do I configure the health checks (formally known as Golden Signals) to consume the json response the Endpoint URL provides?

From “Services”-> your specific service -> “Golden Signals”, select the integration type and then select “Add”.

Now, provide the “Endpoint URL” and any appropriate authentication headers.

(in this example, we manually create a monitor against Datadog for demonstration purposes only; if you have a Datadog signal, we suggest using that specific integration).

Using “Test Request”, the Success Evaluation fields open up.  Configure appropriate values for the http status code, request timeout, and what values to verify in the json response.  The Authenticated response is shown above this configuration to serve as a sample response.

After running a “Test Evaluation”, be sure to “Save” the configuration.

How their Health Checks (formally known as Golden Signals) will work at scale with the Private Network Integration Agent/can they share the net-integration agent, or will they need one for each gremlin team

You only need to deploy a single Integration Agent for your entire Gremlin company. While the agent must be authenticated using a team ID and credentials, every team within your company will be able to use the same agent instance. If you want to restrict the URLs that teams can access using the agent, you can create an allowlist.

Further details here.