I get this error when I try to access the index.php webpage.

PHP Fatal error: Call to undefined function plugin_listener() in /var/www/Graphite-Tattle/plugins/irccat_plugin.php on line 3

When a metric breaches one of the thresholds, an alert is triggered. When a metric that has been triggered falls back below the threshold, an "all-clear" message is sent out.

I have seen these "all-clear" messages sent out when they should not have been, simply due to the fact that the stats had not yet been published for that metric.

When this happens, the JSON will contain a Null value as the very last entry.

I'm about to start using Graphite. After skimming through their docs in search for dashboards I opened about a dozen tabs. I don't want to install and configure all these apps just to see how they look like. Please, add screen shots to README.

The plug-in is able to connect only with smtp over ssl.

Tried to do some debugging on my own, but could not get Flourish to respect the $debug = true parameter. Right now we have Tattle up and running, but none of the alert thresholds are working even though they are 1,000% higher than my error value.

This is what I see on the alerts page.

There are currently no Alerts based on your subscriptions. Smile, looks like everything is happy!

Here is an image of the graph.

http://imgur.com/5ajxr

Any advice to help debug this would be appreciated.

Graphite allows you to ask for something like system.*.cpu_load' and get back an array of data for everything that matches that pattern --- you get back a json array forsystem.host_a.cpu_load' and `system.host_b.cpu_load', etc.

For our use, we'd like to be able to get alerts for any of those wildcarded metrics going over the limit, e.g. "Any system load over 3" or whatnot. Right now, Tattle accepts that as a metric, but only does notification for the first metric that happens to be reported (Check::getResultValue explicitly looks at $data[0]).

Setting aside the future issue of things like "I want notified for all system's cpu load over 3, except big_host, where I don't care until it reaches 10, or test_host, where I don't care at all", I'd like to figure out the sanest way of handling this. Has anyone else thought of this, or wanted this? I've got a few ideas on how to handle the simple case, but I wanted to check with other folks before going too far down that road.

this is an issue only for non-english users, but there are plenty of them around me (incl. myself) :-)

Just pulled the newest source, ran the sql file. Got the following error when I clicked the register button:

Uncaught fProgrammerException
-----------------------------
{doc_root}/tattle/user.php(41): fActiveRecord->__construct()
{doc_root}/tattle/inc/flourish/fActiveRecord.php(1115): User->configure()
{doc_root}/tattle/inc/classes/User.php(6): fORMRelated::setOrderBys(Object(User), 'Subscription', Array)
{doc_root}/tattle/inc/flourish/fORMRelated.php(1343): fORMSchema::getRouteName(Object(fSchema), 'users', 'subscriptions', NULL, '*-to-many')
{doc_root}/tattle/inc/flourish/fORMSchema.php(181)
The table users is not in a *-to-many relationship with the table subscriptions

My DB is forced to use MyISAM. I ran a previous build a few days ago, on MyISAM (different server) and that seemed to work just fine.

Is this a DB issue or a code issue related to registration?

EDIT:

Looking at the Flourish docs, it seems like it wants InnoDB regardless, but will function on MyISAM... but sounds like the programmer has to do more work. I'm very unfamiliar with Flourish.

Hi, I'm a new user and I'm having troubles bringing up Graphite-Tattle with Nginx. Any chance you post an example configuration on how to accomplish that?

On the dashboard edit page, there's a "delete" button for each graph. Clicking this brings you to a blank page and the graph doesn't actually get deleted.

Todo : Need to add pagination to alerts pages. Both front page and individual alert view

Event though the password box has an asterisk, indicating that it is required, it is possible to set someone's password to an empty string.

I was just trying to change their email address, and did not know their password.

I would recommend not updating their password when it is left blank.

We want to add a way to reorder the graphs in a dashboard. When Tattle gets all the graphs, they are already ordered by "weight" ascending. But when two graphs have the same weight, those two are ordered by "graph_id" ascending. That leads to my point : what is the use of the column "Weight" for a graph ? And is this column really usefull ?

We want to add up and down arrows in the graphs list which are more user-friendly. We may also remove the select box in the "add_edit_graph.php" file and rename the "weight" column in "order" or another more explicit word than "weight" for this use.

Does anybody use this column for another thing than sorting the graphs ?

I originally tried running processor.php in root's cron job like so:
*/5 * * * * php /var/www/html/Graphite-Tattle/processor.php
but I got path errors.

So I made a change in /inc/init.php because getcwd will not always set the root correctly unless you are currently in the Graphite-Tattle directory.
I replaced
define('TATTLE_ROOT', str_replace(array('ajax'),'',getcwd()));
with:
define('TATTLE_ROOT', str_replace(array('ajax'),'',dirname($_SERVER['SCRIPT_FILENAME'])));

one caveat is I am using the f80 fork of Graphite-Tattle so the next bit may not be relevant.

I tried running the cron so it appeared to be done by the apache user and it seems to succeed (writes to log, sets alerts, etc.).
*/5 * * * * su -s /bin/sh -c "php /var/www/html/Graphite-Tattle/processor.php -debug >> /var/www/html/Graphite-Tattle/logs/debug.log 2>> /var/www/html/Graphite-Tattle/logs/errors.log" apache

But for some reason I would not receive email notifications. I checked logs, etc. and everything seems fine.

However, I had noticed if I hit processor.php from the web client it worked. So I decided to try this:
*/5 * * * * curl 127.0.0.1/Graphite-Tattle/processor.php?debug=true >> /var/www/html/Graphite-Tattle/logs/debug.log
which works.

I just wanted to see if anyone had any better advice on how to run processor.php. I was thinking I may add in my curl command crontab to the README as a pull request if you like.

Might make sense to move to a json or other formatted database column for storing the graph details vs. having to create a new column per tweak as we constantly add new features.

Might make sense to have an advanced text box on the graph creation page that is just a free form box to add any extra graph related features that aren't built into Tattle yet.

Currently if you have a metric that is reported only semi-frequently (i.e. a job that runs every 1 or 2 minutes) you'll get a lot of nulls back from graphite which throw off the averages and medians. You can wrap the target in a keepLastValue() function but it can still end up with nulls at the beginning of the results. We should probably ignore nulls, but we need to handle the case of all results being nulls.

It'd also be nice to have an option to explicitly alert on all the values being null (i.e. if my job that sends data to graphite is broken, alert on the fact that graphite is not getting any data).

This could be a nice feature to have, for my team we have some checks we usually shut off overnight, or different thresholds for overnight when batch jobs are running. Having schedule based checks would allow for either of those (in the latter case you'd need two separate checks but it's probably the simplest method).

Hi All,

Here my requirement is Integrated BML(Bill Me Late) Hosted page to my existing E-commerce project.

1. Now we are using merchant site only for BML.
2. Instead of this page i want to Integrated BML Hosted page, including user data like address info., etc.
3. How to give link to merchant page to BML secure hosted page.(now we using "click here it's go to merchant page" exactly here i want to go hosted BML secure page.)
4. I seen some Hosted Order Page In Cyber Source website like below link. But i'm not understood much more.
   http://www.cybersource.com/developers/develop/integration_methods/hosted_order_page/

Please help me need full and share your valuable thoughts.

Thanku
All

My alert tests for a metrics to be above 30 for warning status and above 50 for error status. It's consistently between 0 and ~15 most of the time, but I get an alert every valid interval with status '0.0'.

Some metrics have very high spikes which skew the graphs. It would be nice to have a yAxis Max at the graph level to fix this.

Graphite uses the following url attribute to add this:
&yMinLeft=

This happens when you use a target with alias(foo.target,"bar").

prepareTarget turns this into alias(foo.target,"bar"). That's fine for HTML but not for a URL.

solution is probably to have an "encodeTarget" method that uses urlencode.

Hi;

I can not see active alarm

and get the error log

PHP Fatal error: Call to a member function tossIfNoRows() on a non-object in /home/graphite-tattle/inc/views/index.php on line 6, referer: http://graphite-tattle.example.net/user.php

there is a solution for it?

Please add a time based offset for the values which are requestet from graphite.
Usecase:
we calculate different business KPIs/metrics in hadoop with a interval of 3, 6, 12 and 24 hours.
so we cant use checks based on last x values or minutes.

Add the ability to set the timeframe of the graph when creating it.

Example url format:
&from=-2hours

inc/Classes/Graph.php : drawGraph()

I receive rather often false positive for threshold alerts when the alert target contains a complex formula such as diffSeries(sum(foo.bar..baz),sum(foo.bar..foo)), probably because the alert is evaluated on last non null value whereas all metrics are not recorded at the same second within graphite sample interval e.g. if many foo.bar.*.foo metrics are recorded the second before current minute, the diffSeries result will be higher than expected.

There are a few changes we need to make and each one is most likely going to require a database schema update.

Figure out a good way to iteratively update the database without borking any data.

The graphite documentation specifies that the movingAverage function operates on samples, not minutes. However, on a 24-hour graph, Graphite condenses samples into 1-minute buckets, leading to the confusion.

For me, this means that, if I choose "3", Graphite averages the last 30 seconds (since my samples are kept at the default resolution of 10 seconds).

This is confusing, since the UI mentions "minutes", which is only true for the 24-hour graph; whereas the actual polling of the data operates on a much smaller timescale and performs the moving average on a sample-by-sample basis.

See:
http://graphite.readthedocs.org/en/1.0/functions.html

e.g. target alias(foo,'bar') works but target alias(foo,"bar") is truncated to alias(foo,

for both dashboards lines and checks

We are using graphite to track an error bucket of sorts.

We need to get alerts anytime this bucket has more than zero items in in.

There are several issues with Graphite-Tattle that prevent us from doing this effectively:

1. We are unable to turn off moving averages (aside from using a single sample).
2. We are unable to tell Graphite-Tattle to look at a shorter window than 1 day. When looking at a 1 day window, the values are averaged.
3. We don't want separate error or warning thresholds here (but we can use 0.5 for warnings as a workaround).

Thoughts?

The "description" column actually shows Area, and the "background color" column shows the Description, and there is no background color option to display in its place. Need to fix the ordering and naming of the table headers.

In order to quickly eliminate email issues while testing after installation, it would be great to be able to sent a test email

I'm working on a few Tattle enhancement and would like to upgrade to last Bootstrap release.

Does anybody disagree with such change ?

I prefer to ask before because there is a pretty large version gap and therefore I wonder if there is a good reason for still using v1.3.0 nowadays.

Thanks.

We sometimes have errors that last only for a few seconds (CPU usage above the error threshold, for example). We would like to be notified of these issues, even if they aren't still in error when processor.php is being called.

Rather than checking the last value, we would prefer if Graphite-Tattle checked all values since the last time processor.php was run.

This is related to #33.

I use Graphite to store temperatures from sensors once every 15min. As I understood, the "samples" setting is used for 2 purposes :

• how many samples the moving average is based upon
• how far ago processor.php searches for data in the graphite render request using the "from" parameter

I think the two purposes should be considered separately. With my 15min retention frequency, I must set "samples" to at least 15 to be sure to find data to check. But therefore I'm stuck with a moving average of 15 samples, that I don't want.