Full logs archive is slow, issues many queries, and has an HTTP error

[edongashi]

On page load I see this query repeating lots of times

The page is slow to load.

At the end I got this message in the console.

Maybe make the link download the file instead of opening it directly? A raw text file of that size is useless for browsers anyway...

[edongashi]

It crashed my browser. We definitely need to make this a downloadable file by adding the appropriate headers.

[edongashi]

I suppose the queries are stream chunks. In ecto it says it takes 500 entries on one run. Could be made a little higher I think.

[sebastian]

I approve of making it such that the browser downloads a file, using larger chunks, and making it look more like a download button, or make it clear from the link that it's a way to download the logs.

[sebastian]

Also the link doesn't take the filters into account. You always end up with the full log. Maybe that's ok, and should be clear from the "full archive"...

[edongashi]

I approve of making it such that the browser downloads a file, using larger chunks, and making it look more like a download button, or make it clear from the link that it's a way to download the logs.

Yeah, we can make it a button and add a download/file icon.

The chunk size wasn't at fault, the whole file was being loaded in memory due to a bug, and the size was 60 MB which is way too much for the browser to handle.

[edongashi]

Also the link doesn't take the filters into account. You always end up with the full log. Maybe that's ok, and should be clear from the "full archive"...

I say we consider this a different issue.

[sebastian]

I say we consider this a different issue.

And I suggest we permanently ignore it :)