You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the problem/bug
I have a moderately busy routing node, and have around ~500k routings over the last year.
With 0.13.24, a Year aggregate of forwards has been added to the home screen and the forwards screen.
This takes about 6 minutes of processing time on my server, after which a 126MB of GraphQL is sent back to my browser for another 2 minutes of processing time. The memory usage of the Chrome tab balloons to 3.3GB and (I think due to an automatic refresh) after a few more minutes the tab is unresponsive.
The server is an Intel Xeon E-2286G with 64GB memory, storage is SSD. The client is an 8th gen Intel i7, so also not slow by any means.
Expected behavior
In 0.13.23, it took about 5 seconds for the aggregates to show up, with maybe a 2 second process time of the browser. I couldn't extend the graph to anything beyond 7D, as that would blow up the system (or I never waited long enough).
It seems all forwards are requested for the last 365 days, with a GetForwardsList. I would expect that the aggregation of week, month and year is done by the underlying database, not on the browser. My guess is that would speed things up immensely.
In the current state I'm afraid ThunderHub is close to unusable for a decently sized lightning node.
The text was updated successfully, but these errors were encountered:
mutatrum
changed the title
[BUG] Main screen and forwards screen blow up due to new year totals of forwards
[BUG] Main screen and forwards screen unresponsive due to new year totals of forwards
Dec 6, 2023
Describe the problem/bug
I have a moderately busy routing node, and have around ~500k routings over the last year.
With 0.13.24, a Year aggregate of forwards has been added to the home screen and the forwards screen.
This takes about 6 minutes of processing time on my server, after which a 126MB of GraphQL is sent back to my browser for another 2 minutes of processing time. The memory usage of the Chrome tab balloons to 3.3GB and (I think due to an automatic refresh) after a few more minutes the tab is unresponsive.
The server is an Intel Xeon E-2286G with 64GB memory, storage is SSD. The client is an 8th gen Intel i7, so also not slow by any means.
Expected behavior
In 0.13.23, it took about 5 seconds for the aggregates to show up, with maybe a 2 second process time of the browser. I couldn't extend the graph to anything beyond 7D, as that would blow up the system (or I never waited long enough).
It seems all forwards are requested for the last 365 days, with a GetForwardsList. I would expect that the aggregation of week, month and year is done by the underlying database, not on the browser. My guess is that would speed things up immensely.
In the current state I'm afraid ThunderHub is close to unusable for a decently sized lightning node.
The text was updated successfully, but these errors were encountered: