lucjad
Neophyte
NeophyteForum Discussion
I have Microsoft_SQLServer_GlobalPerformance customized and have issue: No Data in Microsoft SQL Server Global performance- MSSQL. Issue of No Data is only tied to few servers. I was prompted by LM Support to update the module. Documentation says that there're risks associated with updating the module. It's customized and monitors more than 100 servers. Is there any way to avoid risks of loosing historical data and loss of custom changes? I don't want the update to affect all the other servers.
lucjadWould I be able to update the custom module and then put the custom scripts back somehow? I mean if I loose these custom changes, can I put those custom scripts back into this specific module? Hope that makes sense.
- Mike_Moniz
Professor
Kinda? What customizations did you make?
You can export the existing DataSource to a JSON file which would retain all the changes you made and you can re-import it so it goes right back to how it was. The trick is that you might lose historical data or customized device/instance-level changes doing that. Depending on the customizations you can have LM upgrade the rest of the module without undoing your customization (like if you only changed AppliesTo), or you can just "upgrade" your customized version by manually implementing the changes LM made (like making the groovy code changes). You can also do the opposite in that you just import the new version then re-modify it to redo your customizations.
- Mike_Moniz
Updating DataSources can be nerve racking and can be little risky although I do think LM does attempt to limit data loss from upgrades and will list possible issues in the release notes and support page.
For MSSQL, there is a note at https://www.logicmonitor.com/support/monitoring/applications-databases/microsoft-sql-server#migration-to-latest-logicmodules about instance names might change, which might cause you to lose history. I wasn't able to locate the release notes that cover the update for Microsoft_SQLServer_GlobalPerformance specifically but perhaps LM support can help you locate that.
If you have a LM Sandbox instance setup, you can try using that to test out the upgrade to determine any affects the new DataSource will do.
If you know how to read Groovy code, you can look at the differences between the current DataSource and the new one to see what it changed and perhaps get an idea what might happen.
You can also try installing the new DataSource with a new name so it runs in parallel with the old one. You might need to use a Sandbox instance to import it, rename it, export to JSON, then import into production instance. If you don't have Sandbox instance, perhaps LM support can export the latest version for you and you can edit the JSON file directly.
