There is no data in the analyzer.

[Mikhail Petrov]

Hello!

I wrote a perl script that sends data to the skyline every minute.

http://pastebin.com/ixCVgEgL

Service horizont receives these data.

(I've added logging in script)

.......

.......

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.032']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.003']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.022']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.011']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.221']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057240', '0.002']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057241', '0.096']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057241', '0.081']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057241', '0.083']

2013-10-29 14:34:29 :: 5455 :: get some metric on ['nginx_response_time', '1383057241', '0.003']

.......

......

I set FULL_DURATION = 90.

But in /var/log/skyline/analyzer.log I see the following ( yes, added logging )

---

        # Discover assigned metrics

        keys_per_processor = int(ceil(float(len(unique_metrics)) / float(settings.ANALYZER_PROCESSES)))

        logger.info(unique_metrics);

---

---

                unpacker.feed(raw_series)

                logger.info(unpacker.unpack())

                timeseries = list(unpacker)

---

2013-10-29 14:34:35 :: 10423 :: WARNING: skyline is set for more cores than needed.

2013-10-29 14:34:35 :: 28194 :: ['metrics.nginx_response_time']

2013-10-29 14:34:35 :: 28194 :: 1383054061 <-- strange value, i think.

2013-10-29 14:34:35 :: 10423 :: seconds to run    :: 0.28

2013-10-29 14:34:35 :: 10423 :: total metrics     :: 1

2013-10-29 14:34:35 :: 10423 :: total analyzed    :: 0

2013-10-29 14:34:35 :: 10423 :: total anomalies   :: 0

2013-10-29 14:34:35 :: 10423 :: exception stats   :: {'DeletedByRoomba': 1}

2013-10-29 14:34:35 :: 10423 :: anomaly breakdown :: {}

2013-10-29 14:34:35 :: 10423 :: sleeping due to low run time...

What am I doing wrong?

Thank you for advice!

[Abe Stanway]

Can you paste the full logs please, including the spin up? Can also you make sure you have the most recent release of Skyline?

-

Pardon the brevity

--
You received this message because you are subscribed to the Google Groups "skyline-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to skyline-dev...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Mikhail Petrov]

Logs http://yadi.sk/d/mZ7Te3B7BnTwJ

Before starting I performed git pull

вторник, 29 октября 2013 г., 15:48:19 UTC+4 пользователь Mikhail Petrov написал:

[Abe Stanway]

I'm not sure I understand. You added your own logging to horizon? That value looks like a unix timestamp to me. Why is it strange that it's there if you are logging for it?

--

You received this message because you are subscribed to the Google Groups "skyline-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to skyline-dev...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Abe Stanway

abe.is

[Mikhail Petrov]

Yes, I added own logging.

                    metric = unpackb(data)

                    logger.info('get some metric on %s' % metric)

                    chunk.append(metric)

in listen.py

Oh, sorry, the main question was, what I see exception stats :: {'DeletedByRoomba': 1} and don't understand why.

I'am read https://github.com/etsy/skyline/wiki/Getting-Data-Into-Skyline

Horizon also accepts metrics in the form of messagepack encoded strings over UDP, on port 2025. The format is <metric name> <timestamp> <value>. Simply encode your metrics as messagepack and send them on their way.

And send data in a specified format:

my $packed = $mp->pack( [ 'nginx_response_time', $time, $5 ] );

Where time is unixtime

Maybe I'm wrong.

What is the correct format for the timestamp?

Thank you!

вторник, 29 октября 2013 г., 15:48:19 UTC+4 пользователь Mikhail Petrov написал:

Hello!

[Abe Stanway]

Roomba is supposed to delete metrics that get old. I think the reason it's showing a 1 there is a minor bug. As for your timestamps, they look fine.

--

You received this message because you are subscribed to the Google Groups "skyline-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to skyline-dev...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Abe Stanway

abe.is

[Mikhail Petrov]

I found a mistake in a perl script.

It is written in the wiki https://github.com/etsy/skyline/wiki/Getting-Data-Into-Skyline

UDP messagepack

Horizon also accepts metrics in the form of messagepack encoded strings over UDP, on port 2025. The format is<metric name> <timestamp> <value>. Simply encode your metrics as messagepack and send them on their way.

But send data should thus:

my $dat = [$metric,[$time,$time_response]];

my $packed   = $mp->pack($dat);

Thank you.

вторник, 29 октября 2013 г., 15:48:19 UTC+4 пользователь Mikhail Petrov написал:

Hello!

[Naveen Voruganti]

Any success with seeing metrics on skyline, I do not see them.