Alert History as Dataset

In each of the individual alert, we can view its Alert History as a line graph or as a chronological list of trigger event. (ref

We create alerts to monitor our system health and want to build our KPI around how often a given alert is triggered. Is there a way we can get Alert History log as a dataset so we can build a chart similar to that shown in Alert History screen?


  • GrantSmith
    GrantSmith Indiana 🔴

    Hi @lwang

    Currently there isn't an a supported method (connector or api) to get alert history as a dataset. You might be able to work with your CSM / Domo to see if they could inject that information into your instance as a dataset.

  • Thanks @GrantSmith. Have reached out to our CSM/Domo Support and waiting for reply.

  • I just wanted to say that I think this is a fantastic idea and happened upon this post today after wanting something similar.

  • jaeW_at_Onyx
    jaeW_at_Onyx Budapest / Portland, OR 🟤

    @JasonAltenburg make sure to put the idea in the Idea Exchange. I promise to upvote. :P

  • @GrantSmith, @JasonAltenburg, @jaeW_at_Onyx I log this ask in Idea Exchange Let's upvote it to make this happen.

    As for workaround options, I got below response from Domo Support, FYI

    " One workaround would be to use one of the additional actions that can be added to alerts, such as webhooks or tasks.

    Webhooks allow you to send a request to another web-based service when the alert triggers, which you can then handle on that end to generate a list of times when the alert has triggered. This would be limited by whatever services you point the webhook to, so it can be flexible or very inflexible depending on what you send the webhook to, and the capabilities of that platform.

    The other workaround is to use the task creation option in the alert. You can assign a project to the alert that will have a task created when the alert is triggered. You can then run the Domostats dataset using the Tasks report to get a list of the times when the alert triggered and what the data was from them. If you use tasks for other things, this can cause your tasks to get very cluttered, so depending on your use case, this might not be ideal. If you do choose to do this, you will likely need to run the resulting data through an ETL to make it easily usable for building KPI cards.

    You can read more about these actions here: "

  • jaeW_at_Onyx
    jaeW_at_Onyx Budapest / Portland, OR 🟤

    @lwang , all these options are ... ok, but if it were me and i HAD to have an alert, i would use a recursive query to construct a dataset that captures my alert conditions.

    so whatever your 'system health' stat is, generate one row per day of the stat, or generate a dataset of the conditions that triggered the alert then recursively accumulate the history, so you'll know, 'yesterday we had this problem, today we had that problem.' OR 'yesterday our health metric was 93, today our health metric is 55'