Behaviors when multiple agents monitor the same resource

A time series is a sequence of data points taken in successive equally spaced points in time.

A time series is associated with the monitored resource attribute, not the agent:

  • Only one time series is created for that resource URI (including the attribute), regardless of how many agents are monitoring the same resource.

  • The time series remains in the time series database as long as there is at least one agent monitoring that URI. The entire time series in the time series database is deleted only when all the agents monitoring that URI are deleted.

If the agents have different conditions or threshold values, actions (such as alert generation) are taken for each condition met, but the time series being used is the same for all agents. As a result, an agent might evaluate conditions at the time it is enabled based on data that existed prior to that time, such as if the condition aggregates data over time.

For example, consider the following scenario:

  1. You create two agents, both of which are disabled:

    • Agent A1 monitors CPU utilization.

    • Agent A2 monitors CPU utilization and has an action to create an alert when the average over the last 1 minute of CPU utilization is greater than 90%.

  2. You enable agent A1, which starts monitoring CPU utilization.

  3. The CPU utilization is above 90% for 5 minutes.

  4. You enable agent A2.

  5. Agent A2 creates an alert immediately—instead of after one minute—because the time series already shows that the average of the previous minute of CPU utilization is greater than 90%.