What is the purpose if auto:archive if it reads logs from the very beginning every time?

Hello,
on the documentation it is advised that auto archiver is set for websites that tracks large amount of data
but on the same documentatino this is written :

" * Matomo archiving for today’s reports is not incremental: running the archiving several times per day will not lower the memory requirement for weeks, months or yearly archives. Matomo will read all logs for the full day to process a report for that day."

If I run the archiver at 10:00, and then on 11:00, will it read logs from 10:00 to 11:00 and add it to the reports or it will start from 00:00 to xx every time you run it ?

is equivalent to:

thanks you for the fast reply.

if I run the archiver at 10:00, and I request some metrics for today at 10:25, it will re-read all logs from 00:00 to 10:25
How can this help speed up the report requests?

In the doc you mentioned:

you can read:

If your website has more than a few hundreds visits per day (bravo!), waiting for Matomo to process your data may take a few minutes. The best way to avoid these waiting times is to set up a cron job on your server so that your data is automatically processed every hour.

Thanks for the reply but the mentioned part is not clear for me.

How can it help speed up things in the case I mentioned?
Will it only return values up to 10:00 (time I did the archive) in the period of 10:01-10:59 (before next archive) ?

because if not, it will have to re-read logs from 00:00, which is the part that I don’t get, if it’s going to re-read all logs anyway, why bother doing the auto-archive?

I know I’m missing something, please bear with me, thanks.

auto:archive won’t seed up the compilation itself

It “speeds up” because when you want to access a report, some data need to be compiled.
The CRON (auto:archive) allows the data compilation, periodically.
When the auto:archive is set, reports will use the already compiled data. If the data are not compiled, you’ll have to wait the server compiles them.

that’s it.

Thank you for your reply.

so if I understand this, any reports that from X date to today will always trigger new report hence the auto compilation will not have any use for this case

however if I request anything from X date to yesterday, auto compilation will help access already compiled data since it doesn’t not make sense to recompile old data ?

thank you again for your time!

The daily reports of finished days won’t be re-archived. Once they won’t change (after 12PM, the report won’t change as new data will be included in the report of the following day), they are not rearchived again…
The same for weekly, monthly and yearly reports.
But for other periods (eg. from 2nd of September to 11th of September, AFAIK, no report will be automatically generated. Data for such reports will be compiled on demand (probably with daily archives).
@Lukas, am I right?