Only me is hating the new error’s panel??? I really can’t use it!! One of the principals tools from Newrelic is terrible and for me it doesn’t make any sense. if this continues so canceled my contract and use other tools.
I’ve just had a look at the errors in APM, if that is the one you mean @eoliveira?
It looks ok to me.
It might be more helpful to specify what you are having difficulty with or suggest improvements.
Just change everything.
no one can understand anything, I can not know when there are problems, when there is increased any errors. Does not show servers that are giving error.
My suggestion is that back immediately as it was before, if they want to implement something different, talk to your customers before. ask for suggestions before and not after they made a bonehead.
Honestly, if left as it is, you lost me as a customer.
Just to clarify, I do not work for New Relic.
Unfortunately I did not see the previous errors so I do not understand your frustrations.
yes this has causes some confusion when they made this change mid day , we liked seeing the server that was tossing the errors as well and that has seemed to go away with this new update.
Buts just for you understand, before to have this change, we could see the amount of errors separeted by type and if necessary we could click on error e see wich server was generating this and also the amount by servers in a period of time.
If we had an error the graphic instantly climbed the amount of the error and what was already said, what server or servers which …
now it appears a graphic of Windows defrag… nobody can understand it! really! it is terrible… doesn’t help nothing.!!
I can not identify which server is in trouble, this update was a mistake, please go back to the previous version.
I’m looking into this now for you. I’ll be back with some information shortly.
Unfortunately, we are currently only using one application name per server, and separating our apps that way. I say unfortunately, because that means I can’t really judge the exact problem you’re having. On a whole, though, I really like the new errors interface, because it lets me see exact (ish) timing of each individual error instead of just getting a roundabout number of “200 similar errors in the last 3 minutes”
Can you potentially push the server name as a custom parameter?
On a clustered application the only workaround I can see is to select given node in the “Overview” page before navigating to the “Errors” tab.
Ideally the “Error traces” tab should contain a column for the server name
Do you have a permalink to an application so I can have a look? It’s safe to share permalinks on the forums as only users on your account and NR staff can use them.
other problem… our custom dashboard stopped of works… i have no more informations…
Sorry… i puts information in wrong forum. but it’s real… my custom dashboards are not working anymore… and the graphic bellow is what i want to have back. our error rate
It is good
It is terrible
Ok, some people like but i spoke with some friends that using newrelic and 100% of them hate the new feature.
Ii should be an option so Newrelic will please everyone.
I agree. I would like the old Errors back. Its very frustrating trying to use the new Errors and I haven’t been able to resolve error reports I’m seeing because its hard to tell where they are coming from.
@eoliveira and @gmoy - I see that you both have tickets open. Thanks for getting in touch with us - we’ll work with you in those tickets, so that it’s easier to discuss account settings. If you’re having any trouble accessing the tickets, please let us know!
I absolutely agree with @eoliveira. The new error UI is borderline unusable for us. It used to tell us clearly which servers are having issues, similar errors aggregated by counts and so on.
We haven’t been able to investigate error rate spikes since the new UI. The new UI is very hard to understand even for techies and feels nowhere as useful as before. Is there a way to revert back to the old error UI?!
Hi, a few other wishes:
- When you are looking at Error frequency section, the frequency chart entries for errors is hard to use when you have multiple classes, and have a class that has a significantly high error count, the gradient range makes it so that classes with lower errors are impossible to visually distinguish. Gradients should be per class, not global
It would be nice if errors were filtered as pages were selected/de-selected in the “top 5 errors” graph
Error traces would be much more useful if they included requesting IP address and which server responded to the request. In the example shown above, we were getting pummeled by malformed traffic; it would have been helpful if I could have seen that all this traffic was coming from one place in the errors panel
As others have mentioned, filtering by server name would be very helpful as well
I think “useless” and “terrible” are a bit overstated, but the Errors panel would definitely benefit from some polish.
I also agree. The new errors UI doesn’t feel like an improvement. I wish it was as it used to be, as it was the most used part if NR APM I used. Especially how URLS are no longer recorded for errors. That was SUPER helpful for me to re-create errors locally for debugging.
@jason_t that’s fantastic feedback, thanks you so much for going into that level of detail. As Filterable errors is still an evolving part of the product, I will fill a feature request or two for your ideas. Certainly if our product managers feel that these ideas add a bit of polish to the product they might consider them for the roadmap, so i’ll file that for you now. Thanks for helping us improve the product.
You can filter by host with the new page Jason, try this link from your account. If you want me to walk you through any of it just tag me and ask me to create a ticket and i’ll be glad to help out.