There is one named task chain.

It runs every 5 minutes.. does it's work, and then schedules itself to run
again at the next 5 minute mark (not in 5 minutes from when it ends.. but in
how ever many seconds there are until the next 5 minute mark).

This is the only thing my app does during the day.

So I can see a log entry at 9:30 AM, 9:35 AM, ... all the way until 1:05 PM.

There is a secondary task that this chained task fires off after it finishes
each run.  That task is set to run immediately, and it only runs once.. and
does not chain itself.

This makes it very easy for me to detect if/when my particular app
experiences any weird taskqueue delays.  Since.. the main task should run at
each 5 minute mark.. if it doesn't, I know there is an appengine delay
there.

The secondary task should start exactly after the main chained task finishes
its work.. so i can easily detect delays there as well.

When everything is running smoothly, I can look at my logs and just see
pairs of requests throughout the day, like so:

01-03 12:30PM 01.478 /oneOffTask/go 200 344ms 381cpu_ms 241api_cpu_ms 0kb
AppEngine-Google;
01-03 12:30PM 00.694 /cron/chainedTask 200 794ms 665cpu_ms 408api_cpu_ms 0kb
AppEngine-Google;
01-03 12:25PM 01.690 /oneOffTask/go 200 348ms 405cpu_ms 241api_cpu_ms 0kb
AppEngine-Google;
01-03 12:25PM 00.323 /cron/chainedTask 200 1380ms 774cpu_ms 408api_cpu_ms
0kb AppEngine-Google;

So.. by glancing at the log, I can detect any sudden issues (since the app
only does this scheduled back-end processing there are no other log
entries).. like if any of these log entries showed times of 12:27:11 or
12:33:28.. etc.. I'd need to look closer at the logs to see why the start
time was off.

For my app, nothing like that appears to have happened today.

Now, if you had a heavily used queue.. and it was hitting its tasks per
second limit.. that would give you a delay.


On Mon, Jan 3, 2011 at 5:44 PM, vlad <[email protected]> wrote:

> @Eli   Are running a single task chain or there could be many chains
> executing independently?
> The reason that is important is GAE opportunistically uses *any*
> ready-to-execute task in a  given queue to check whether other are ready.
> What that leads to is real schedule will vary depending on how many tasks
> are scheduled.
>
> Generally I have to agree that not every 0s delay tasks suffers huge
> delays. I wasn't able to discern a pattern so far.
>
>
>
>  --
> You received this message because you are subscribed to the Google Groups
> "Google App Engine" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<google-appengine%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/google-appengine?hl=en.
>

-- 
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en.

Reply via email to