Max memory and restart windows server

[PhamHuyThien]

Hello I am currently having some major problems using it

After starting the system for about 1 hour - 2 hours, RAM usage suddenly increases to 90% and restarts the server
Currently, I have used PM2 and rebooted every hour but the RAM is still at 90%.

My device configuration:
Windows server 2016
RAM: 16GB
CPU: 8 cores
400GB SSD

my source:

async investing(body = {
        type: 'klines/realtime',
        name: '', //https://github.com/DavideViolante/investing-com-api/blob/master/mapping.js
        period: 'P1D, P1W, P1M, P3M, P6M, P1Y, P5Y, MAX',
        interval: 'PT1M, PT5M, PT15M, PT30M, PT1H, PT5H, P1D, P1W, P1M',
        pointcount: '60, 70, 120',
        pptrLaunchOptions: undefined,
    }) {
        let response = await investing(body.name, body.period, body.interval, Number(body.pointcount), body.pptrLaunchOptions);
        return ....
    },

Looking forward to receiving comments.

[PhamHuyThien]

https://github.com/PhamHuyThien/investing-com-api-v2

I forked the project and upgraded to version V2.0
Fix error: RAM memory overflow

[DavideViolante]

Hello, can you elaborate more how did you fix your problem, for future users? Is there a memory leak in my repo?

[PhamHuyThien]

Hello, can you elaborate more how did you fix your problem, for future users? Is there a memory leak in my repo?

Hello author,

Not receiving a response from you, I think I have to solve this problem myself.
According to my investigation, closing the browser every time the API is called causes a big problem, sometimes 'puppeteer' cannot close itself (when making multiple calls in a row) the chromium windows stay active, stopping the project also does not solve the problem
The way I solved it is instead of starting puppeteer continuously I just open a browser and open multiple tabs and end it each time I'm done working.
I tested opening/closing 200 tabs continuously for 3 hours, the number of suspended tabs was very small (+-5 tabs).

Use my version 2 or upgrade your project, better experience.

End :)

[liardoecp]

Hi everyone,

i was wondering if puppeteer is really necessary. I forked this repo and replaced puppeteer with fetch. It's seems works very well and the performance are better.

Here my PR: #141

Benchmark

Code

const run = async () => {
    const iterations = 10
    const responsesTime = []

    for (let index = 0; index < iterations; index++) {
        const start = new Date()
        await investing('1')
        responsesTime.push(new Date() - start)
        console.log(`run ${index + 1}: ${responsesTime[index]}ms`)
    }
    const avg = responsesTime.reduce((acc, responseTime) => acc + responseTime, 0) / responsesTime.length
    console.log(`avg: ${avg}ms`)
}

Results

With puppeteer

run 1: 1132ms
run 2: 3223ms
run 3: 1433ms
run 4: 1612ms
run 5: 1088ms
run 6: 1833ms
run 7: 1367ms
run 8: 1674ms
run 9: 1715ms
run 10: 1540ms
avg: 1661.7ms

With fetch

run 1: 371ms
run 2: 170ms
run 3: 139ms
run 4: 150ms
run 5: 138ms
run 6: 159ms
run 7: 170ms
run 8: 154ms
run 9: 164ms
run 10: 133ms
avg: 174.8ms