Start time of an action

In the “action info” window of an action, we see start time and end time of the action on that particular server. I had some queries regarding this:

Q: Does the start time shows the time at which the server received the UDP ping?
Q: After receiving the UDP ping, If a client is in it’s evaluation loop, can that delay the server’s mechanism to pull action?
Q: The analyses which are set to evaluate once a day, are they evaluated at 12AM or they are evaluated after every 24 hours from the analysis activation time?

The reason I’m asking these questions is because in my infrastructure, UDP is enabled. But after 12 at midnight, the actions start taking long time to complete. They are stuck in “not reported” When I see the action info, I often see a discrepancy between the action issued time and start time of the action on the server, and it seems to be different for different servers. But during normal operation hours, the this discrepancy seems to be minimal.

I have already moved the BigFix dB defrag job to early morning everyday. This performance fiasco has really caused me a headache, would appreciate some insights.

No. UDP is primarily a notification mechanism that informs the client that new content is available. The typical action processing flow is:

UDP Notification → Content Gathering → Relevance/Condition Evaluation → Downloads (if required) → Action Execution (Start Time)

Yes. If the BESClient is busy processing other evaluations or activities, gathering and processing new actions may be delayed. There are several BESClient settings that control evaluation behavior and responsiveness, which can be reviewed and tuned if necessary.

No, analyses configured to evaluate once a day are evaluated based on an interval, not at a fixed clock time such as midnight. There is no setting to specify an exact evaluation time.

The actual evaluation depends on the client's evaluation cycle and workload. For example, a 24-hour evaluation interval might extend to 25 hours or longer if the client is busy.

I would suggest first reviewing the affected actions to ensure there are no execution time restrictions, scheduling conditions, or other constraints causing the delay.

If everything looks fine, check the BESClient logs for frequent GatherHashMV entries, which can indicate that the client is receiving UDP notifications and initiating content gathering. However, this alone does not confirm that every notification is being received successfully.

If UDP communication is working and there are no action constraints, I would recommend enabling the following on one affected endpoint:

  1. BESClient debug logging

  2. BESClient Usage Profiler

These should provide better visibility into what the client is doing during the affected period.

Sometimes, poorly written relevance expressions, faulty custom content, or expensive evaluations can consume significant client processing time and delay other activities. I've seen cases where a problematic relevance expression takes an unusually long time (sometime hours) to evaluate and impacts overall client responsiveness.

The Usage Profiler should help identify such expensive evaluations, while debug logging can help narrow down where the delay is occurring.

If the clients are eventually receiving and executing the actions, I would largely rule out permission-related issues.

Another thing I would suggest is reviewing the BESClient settings, particularly quiet hours, power-related settings, or any other restrictions that might delay action execution.

Also, check whether the affected devices are powered on and reachable throughout the day or only during their operational hours. This is especially worth checking for cloud-hosted devices, which are sometimes scheduled to shut down outside business hours.

Yes, that's normal to some extent and generally not a concern. The Action Issued Time does not mean the action will start at the same time on every endpoint. As highlighted earlier, action execution depends on several factors, including client workload, evaluation cycles, content gathering, downloads, and execution constraints.

However, if the delay is unusually long or consistently occurs during a specific time window, it's worth investigating rather than simply ignoring it.

Hello @vk.khurava ,

Thanks for your inputs. I have checked that there are no action constraints, no scheduled times or no blockers which can contribute to a delay in action status reporting. This performance hit is significant in the 12AM-2 AM window. After this, it is back to normal so I also doubt that there are any analyses which are causing this problem. This is during patching windows so the servers are also up and running. I’m quite speechless while explaining to the patching teams on why BigFix tends to get super unresponsive during these 2 hours only.

In that case, I would suggest checking how the actions are being deployed. Are you using baselines or Multiple Action Groups (MAGs)? Are these actions scheduled or deployed for immediate execution? Also, how many components are included in each action?

If multiple actions or large baselines are deployed simultaneously without proper scheduling or staggering, the BESClient may experience increased workload due to relevance evaluation, downloads, and action execution. Depending on the complexity of the actions and available system resources, this can delay action processing and affect client responsiveness.

I would strongly recommend reviewing the BESClient logs during the affected window (12 AM–2 AM). They should provide useful insights into what the client is processing and where the delays are occurring. Alternatively, enable debug logging and the Usage Profiler, as suggested earlier, to identify any expensive evaluations or other activities keeping the client busy.

These are single actions(a simple powershell or windows script) which are executed immediately, not scheduled.

Thanks for confirming. Since you've ruled out scheduling constraints and endpoint availability, I would suggest focusing on the BESClient logs during the affected window.

As suggested earlier, enable debug logging and the Usage Profiler on a few affected devices. This should help identify whether the delay is occurring during action gathering, evaluation, execution, or reporting. Without these details, we're just guessing at possible causes.

By any chance is there a firewall block which is delaying BES client to default polling interval which is 6hrs. Also worth checking if there is another activity scheduled during same time frame, worth checking action history

Another potential consideration: All Endpoints stop updating for a specific time of the day - #8 by Aram

One way to verify whether or not the UpdateHistoricalCounts process is a contributing factor here is to temporarily turn on FillDB performance logging overnight.

Action history has already been checked…No other actions running. Command polling is already set for 20 minutes. But, on a few servers, the action issue time and action start time has a gap of more than 20 minutes, which is a worrying sign.

Hello Aram,

Update historical count is already set to 0 in my environment.