Error Counts?
Orv opened this issue · comments
Can the error counts be broken down into types of errors? If so that would be helpful. Even more helpful would be a list of nodes associated with that particular type of error.
Good ideas, something needs to be done to give better visibility into the errors (without looking at the server logs).
I think the type of errors will be easier, I think that's stored that in some JSON in the database with the collector statistics, but wasn't added to the UI. As for the node list, I need to finish the reverse DNS lookup. I started on it early on but focused first on getting other features incorporated.
I was mistaken, I wasn't saving the counts where it is readily accessible. But that just means I can implement something that captures the affected nodes so we can get more information.
I refactored the network polling functionality so that it looks up the DNS name for nodes that have problems and returns that information in a more usable structure to the calling function. With that, instead of just storing the name/IP of the nodes with errors in a summary column, I'm going to add a new table for node errors. It will store the node IP, name (if available), error type, and server response.
My initial thought is to add a "failing nodes" section on the overview page with columns for each of the different error types, with each column contains a list of the nodes that failed with that error (in the most recent run). A future enhancement would be a page where past errors can be viewed and the full responses. I'm not sure if there's a good way to display the response on the overview page (although I guess I could always link to a "details" page).
@Orv does that sound like it's along the lines of what you're thinking? I can share a screenshot once I get some of the UI elements put together.
I WOULD like to see a screenshot, thanks.
@Orv here's my first pass at a summary of the node errors on the main page. Just lists the node name/IP grouped by the error type. I plan on adding another page with the details of the response.
Edit: the response details are most useful for HTTP errors and parse errors, the connection and timeout errors typically all look about the same