How can I keep data accurate in Insight?

[#1]

Hello Jira Service Management community! It's time for some more advice for working with Insight. We’ve spoken about what data to include in Insight, we’ve discussed how to structure it, the next step is how on earth do you keep it accurate? Today I'll share some advice on how to maintain your data in Insight. As always, if you have more advice to add or any questions, comment below.

First things first

100% accuracy at all times should be the goal but in reality that may not be possible and that’s okay. As long as the data is sufficiently accurate enough to offer more business value than not having it in the first place, then you’re at net positive. Many CMDB projects can be delayed or even fail as they’re waiting to be ‘perfect’ before go-live. But sometimes it’s not worth the effort as highlighted by this diagram that highlights Forrester’s recommendations on the matter.

But as I mentioned, 100% accuracy is what we should aim for so let’s look at things we can do to keep data up to date. The advice laid out in earlier articles covers how to pick and structure your data to make it easier to maintain in the first place so check that out as well (links at the top of this post).

Run network scanners and integrations as often as you can

Insight Discovery is a network scanner (now free!) that will find and bring in information about your network items to Insight. It’s useful for bringing in data in the first place but it’s also very handy for keeping data up to date. Run it on a schedule so it can find changes and update Insight accordingly. You can even have it trigger alerts on any changes you’re particularly interested in. The same concept applies for integrations to third party tools and importers for databases, CSV files etc.

You do need to find a balance on how often you should run Insight Discovery, importers, and integrations. Too infrequently and Insight will be long out of date. Too frequently and and it could consume a lot of resources depending how many objects you’re dealing with. Some users run integrations every hour, others may run once a week or even on demand.

We recommend running it as often as you can during quiet times. Take a look at how often you think data will change and the importance of that data to determine how often you need to schedule it to run. With Insight Discovery you can have different scanning patterns run at different frequencies to lower the resources required to keep Insight as up to date as possible.

Automate

You can use custom field data in Jira issues and post functions to keep Insight up to date based on what’s going on in your issues. Let’s take a look at an example.

Someone submits an issue for their laptop and, in Insight, they are registered as the owner of that laptop object. This will pull in the laptop details into the Jira issue. If the solution is to assign them a new laptop from the inventory, which the agent can do via a custom field in a transition screen, then the Insight database needs to be updated.

You can use post functions to update attributes and do the following:

• Remove the requester as the owner of the old laptop

• Set the status of the old laptop as ‘out of service'

• Set the requester as the owner of the new laptop

• Set the status of the new laptop as ‘in service’

Using these kind of automations takes a huge load off the agents as they don’t need to remember to update anything. It’s all taken care of. Similar rules can be brought in for change requests, for setting the status of objects during an incident, and for onboarding requests where the hiring manager can fill out some details of the new employee and post functions can take those details and create a new object for that employee.

Automate more

If you are using issue text fields to enter or update data in Insight, or if you enter objects into Insight manually on occasion, there may be times when the data becomes a bit messy. In these cases, automations can help again!

As an example, server names. Usually these will be standardised and might be easy to mistype. You can create automation rules that trigger when an object is created or updated of the type server to ensure the name meets the naming convention and flag it if it spots an error.

Keeping Insight tidy and standarized is going to help with reporting and queries, make it more readable, and make maintaining it easier as you’re not constantly trying to standardise things.

While automating as much as you can and running your integrations and importers regularly will help, you will still need to do periodic audits to catch anything. You can even use Insight automation rules to send an email reminder to someone to do an audit of data if nothing has changed in the object schema in a while. 

If you find something has become inaccurate, either during an audit or during the workflow of a Jira issue, always look to understand why and if there’s an automation rule that could be implemented to avoid it happening again.

Using these tips you can reduce the burden of admin work and focus more on using the data to close issues faster and understand dependencies.

Do you use automations for keeping Insight up to date? If so post below to give inspiration to others just starting on their Insight journey!