Messaging Monitoring Dashboard

Why should I monitor the agents? And why use this interface?

As you can already monitor your data (article here), the idea is to quickly visualize key information about connectors and parsers in a graphical way.

The monitoring interface helps quickly identify if actions are required for your messaging system to be as effective as you want it to be. 

In the current version, you have information about the status and the queue/dequeue system. The upcoming version will offer additional KPIs to help you efficiently manage your agents.

In addition to the messaging monitoring page you can use the related datasets available in the Data Explorer (Monitoring).

Introducing the Monitoring Page: Filters and General Information

This page is the first one of the new version of the messaging module. It contains 3 monitoring widgets yet and more information will be added soon.

(1) You can refine the data displayed in widgets using the filters available in the filter panel, accessible by clicking on the tab on the left of your screen. Thanks to it, you can focus on specific connectors selected according to their LPar, tags, type ou status. 

Select the expected criteria with the dropdown lists on the panel.

(2) The page is updated every 5 minutes, as are all widget unless otherwise indicated. You can see the last update time at the top of the page. You can refresh it manually if needed thanks to the dedicated button.

Introducing the widgets

The first three released widgets help you quickly monitor the health of your (de)queue system and the status of your connectors and parsers.

Let's describe them in more detail:

Queue Current Ratio

Both stored and processed messages are monitored, allowing you to see their amount in the widget as well as their evolution, providing a clear visualization of the overall trend.

To aid in understanding and analyzing the trend, we calculate a ratio based on the number of stored and processed messages over the last five minutes. The ratio is determined by dividing the number of processed messages by the number of stored messages.

• If the ratio is 1 or higher, the system is performing well: it processes at least as many messages as it receives (or it is catching up).
• If the ratio is between 0.9 and 0.7, the system is functioning adequately. 
  • The key point of interest is the trend of message numbers. You should ensure that the number of processed messages does not decrease significantly and that the number of stored messages does not increase substantially.
• If the ratio is 0.6 or lower, it is considered critical. 
  • This indicates that too many messages are being stored compared to the number of messages processed, creating a bottleneck. This condition is displayed in red to draw your attention.
  • What to do in this case ?
    The problem probably lies with the parsers. Go to table and check the followings:
  • if your data is processed or not, if not try to stop and relaunch the parser
  • if your parser has the status "error", also try to stop and relaunch the parser
  • if the parser runs correctly and that the "Usage Value" shows a percentage of 90%, you have to add a new instance in "Edit":
    • either manually by entering a higher number of instances in the related field
    • or you can also activate the autoscale functionality

Please note that this widget is not affected by filters.

N/A or NA stands for Nan Applicable

NaN stands for Not a Number

Connectors Status

The essence of this widget is to help you quickly identify unexpected errors in your connectors. It displays the status of all your connectors:

• Running: the connector is currently retrieving messages
• Ready: the connector has been started but there are no messages to retrieve
• Starting: a transitional state lasting a few seconds, indicating that the connector has been launched (the next state will be either "running", "ready" or "error")
• Stopping: this a transitional state lasting a few seconds, indicating that the connector has been stopped but is still retrieving the last messages in the queue (the next state will be "stopped")
• Stopped: the connector is inactive (it has been manually stopped or the machine is down)
• Error: the connector has encountered an error
• Warning: at least one instance of the connector is in "error", not all the instances

The number of errors is displayed at the top of the widget. 

When there is at least one error, this is highlighted in red for quick identification. 

A link is provided to access the connectors table sorted by error status, allowing you to inspect the logs and take action (manually restart the connector or set up auto-start to automatically avoid service interruption).

Parsers status

This widget provides the exact same information as the connectors' widget but is applied to the parsers!