fix(qbittorrent): Improve webclient performance

[brettpetch]

improve performance on large queues. no negative impacts.

Description

Modify sizes & timeouts to improve performance with large amounts of torrents.

Fixes issues:

• Un-tracked issue... Large torrent queues become unresponsive in queues ~5500 torrents.

Proposed Changes:

• Modified nginx conf for qbit, queues above 5500 torrents would have issues with responsiveness.

Change Categories

• Bug fix

Checklist

• [x] Docs have been made OR are not necessary
  • PR link:
• [x] Changes to panel have been made OR are not necessary
  • PR link:
• [x] Code is formatted (See more)
• [x] Shellcheck isn't screaming (See more)
• [x] Prints to terminal are handled (See more)
• [x] I have commented my code, particularly in hard-to-understand areas

Test scenarios

Architectures

	amd64	armhf	arm64	Unspecified
Focal		✅	✅
Bionic
Buster
Stretch	✅
Raspbian	⚫️		⚫️	⚫️

✅❎ Passed

✅

🛠🛠 TODO

None afaik

❌❌ Currently failing

None afaik

[flying-sausages]

Did you actually test this out on all those arches and OSs? :p

[brettpetch]

Did you actually test this out on all those arches and OSs? :p

mb. Testing was split between stretch and ubuntu focal for ARM devices. service nginx configtest passed on all counts.

[brettpetch]

Should be set now. Testing still required-- issues could exist, but internet has been unstable. Next possible test will on monday morning EST.

[flying-sausages]

tests pls

[brettpetch]

testing now.

[flying-sausages]

testing now.

how'd it go? 😛

[brettpetch]

hey, just looked back into this. The only part that really mattered from it was the following:

    client_max_body_size 24M;
    client_body_buffer_size 128k;    
    fastcgi_buffers 8 16k;
    fastcgi_buffer_size 32k;

[flying-sausages]

Are we ok with forcing the updater to overwrite the whole qbit config? I am personally fine with that

[brettpetch]

I am fine with that as well. It’s unlikely people mod it too much.

/b

On Jul 15, 2021, at 4:49 PM, flying-sausages @.***> wrote:

 Are we ok with forcing the updater to overwrite the whole qbit config? I am personally fine with that

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[brettpetch]

Are we ok with forcing the updater to overwrite the whole qbit config? I am personally fine with that

@flying-sausages here's my current thoughts on this, we can ship it, but what if we used sed to change only the affected lines in an updater?

/b

[luzpaz]

Any more progress on this ?