[Metrics] What should we do with requests to save data points without a timestamp
by Thomas Segismont
Hi everyone,
While doing some testing for Grafana/Metrics integration, I found that
if you don't supply a timestamp when you send data points to Metrics,
they all get stored with timesptamp=0. This not right.
In such situations, Influx saves the point with the server timestamp. We
could also reply with 400 Bad Request. My preference goes to the latter.
Any thoughts?
I'll file a JIRA when we agree on a solution.
Regards,
Thomas
9 years, 3 months
[GSoC] Hawkular Android Client: Weekly Report #2
by Artur Dryomov
Hi everyone,
This year I am working on the Hawkular Android application as part of the
Google Summer of Code 2015 program.
I’ve spent the last week on the application development and writing a
script which will allow me to populate a Hawkular instance with sample data.
The application itself [1] at this point can use a custom server URL. After
pressing a button you will see an OAuth dialog which will allow you to
login using your Hawkular credentials. This step will automatically
redirect you to the resource types screen. AeroGear did a great job with
the authorization process — seems like all necessary headers are set
without any additional cost from my side. Regarding the fun part — the
application got a fancy theme in red and orange colors based on the
official logo. The icon is there as well. You can take a look at some
screenshots [2].
With the help from various Hawkular modules source code and useful links
from Jiri Kremser I’ve created a gist with HTTPie [3] commands [4]. My goal
was to populate a server from the ground up — from tenants to metric data.
Unfortunately I’m having some issues with it.
⋅ First. I am able to create a tenant but if I try to create an environment
on it — the response gives me 403 Forbidden. At the same time creating an
environment for a tenant from "/inventory/tenant" is flawless. Maybe it is
somehow related to metrics-inventory communication. Inventory has a single
tenant, metrics can have multiple. I see no ability to set an active tenant
for the inventory using the API.
⋅ Second. Seems like I’m adding data to metrics the wrong way. It is
created without any issues, but I cannot see it from the query.
This week I would like work on those areas implementing rest of Inventory
screens for the application and fixing and expanding the script to get
something to display at the application :-) If someone has a real-world
example of multiple resources, metrics and metrics-related data — it will
be a great help.
Have a nice week!
Artur.
[1]: https://github.com/hawkular/android-client
[2]: https://github.com/hawkular/android-client/pull/5
[3]: http://httpie.org/
[4]: https://gist.github.com/ming13/e7ebab69028adaa7dea4
9 years, 3 months
Which logging categories can be hidden by default?
by Peter Palaga
Hi *,
I hope I am not the only one who finds that there is a lot of
uninteresting messages in the log which can be made hidden by default.
I propose to add the following lines to standalone xml files for both
dev and production profiles:
<logger category="org.apache.cassandra">
<level name="WARN"/>
</logger>
<logger category="org.jboss.as.ejb3">
<level name="WARN"/>
</logger>
<logger category="liquibase">
<level name="WARN"/>
</logger>
<logger category="org.ocpsoft.rewrite">
<level name="WARN"/>
</logger>
<logger category="com.datastax.driver">
<level name="WARN"/>
</logger>
Is this too much hiding or maybe somebody has even more candidates for
suppressing?
Thanks,
Peter
9 years, 3 months
Mocking HTTP Server
by Peter Palaga
Hi *,
it happened today: http://perfcake.org/ - one of the sites we unit-test
Pinger against is down and therefore the tests are failing.
Can anybody recommend a solution for mocking an HTTP Server?
http://wiremock.org/ is the first result on Google and I am going to
have a look at it.
Thanks,
Peter
9 years, 3 months
Reasons not to use org.jboss.logging.annotations.Cause at all?
by Peter Palaga
Hi *,
I am just improving the log output in Pinger and Availability Creator
[1] and I found out that one can define a LogMessage containing all the
details of a thrown exception (incl error message and stack trace) like
this:
@LogMessage(level = Logger.Level.ERROR)
@Message(id = 5007, value = "Could not send a message to the bus")
void eCouldNotSendMessage(@Cause Throwable e);
Indeed, messages like this can be seen in Bus, Agent and Inventory, but
nowhere else.
Are there some reasons for *not* using this @Cause message style at all?
It is clear that not every thrown exception needs to clutter the log
with its stack trace. But generally, things that are supposed to work in
all cases and situations (like sending messages to bus) should be logged
incl. stack traces, should they not?
Thanks,
Peter
[1] https://issues.jboss.org/browse/HAWKULAR-309
9 years, 3 months
Alpha1 is out, en route for Alpha2, Alpha3...
by Thomas Heute
Thank you for the hard work on Alpha 1 release.
The next prioritized topics would be:
- More Wildfly monitoring. The WF agent expose quite some metrics now,
only a few of those are currently displayed in the UI
In particular:
- HAWKULAR-217 [1] App Server Detail - JVM Metrics Tab
- HAWKULAR-218 [2] App Server Detail - Web Metrics Tab
- HAWKULAR-220 [3] App Server Detail - Deployments Metrics Tab
- Some Wildfly operations (part of HAWKULAR-220)
- "HAWKULAR-130 [4] Multiple alerts for a single issue" is close to my
heart as I find it annoying --> It doesn't have to be solved by the UI
team AFAIK, the UI shouldn't change (much).
- More integration tests (move them ?)
- ...
Beside the UI, there are quite a few things to add or start to think
about if you look at the mockups [5]
- inventory filtering/searching if that's not yet done, this was not
implemented as part of Alpha1 UI AFAIK
- actions on deployments (Enable/disable/redeploy/(add/remove)) ->
this is a new area, we should start now, it may not make it (all) for
Alpha 2
- some metrics aggregation (Active sessions for all webapps of a
single WF server for instance)
- ...
PS: Mockups attached to JIRAs are not correct
Thomas
[1] https://issues.jboss.org/browse/HAWKULAR-217
[2] https://issues.jboss.org/browse/HAWKULAR-218
[3] https://issues.jboss.org/browse/HAWKULAR-220
[4] https://issues.jboss.org/browse/HAWKULAR-130
[5]
https://docs.google.com/document/d/1kLOGUyajyMsZuJoUxrLJkOzIyQb1V_r4JHj99...
(Unfortunately this doc which adds more details is not accessible
outside Red Hat employees)
9 years, 3 months
Hawkular-pluggable data processors for metrics: Weekly Report
by Aakarsh Agarwal
Hi all,
I am currently working on the Hawkular Metrics repository as part of the Google Summer of Code 2015 program and this is to update the progress of my project.
I spent the last week familiarizing myself with the existing Hawkular Metrics repo on github and some of its code. As of now, I have created a new git repo on my github profile which intends to implement a simple plugin interface using 5 basic statistical algorithms. These are -
Minimum of the entered data set.
Maximum of the entered data set.
Average of the entered data set.
Mode of the entered data set.
Standard Deviation of the entered data set.
The link to this git repo is - https://github.com/Akki5/hawkular_plugin
This is just initial draft for the interface to be implemented. Here, StatisticalAlgo.java is the interface and other named java classes are the classes implementing the interface. PluginDemo.java is the test class. At the moment, everything in the code is hardcoded and static in nature which will later be converted for dynamic usage.
The next step requires following listed work to be done -
Use maven to build the project.
Use Classloader for all classes implementing StatisticalAlgo.
Require input from user as to which plugin is to be used.
Regards,
Aakarsh Agarwal
9 years, 3 months
On the way to the M1 release ...
by Heiko W.Rupp
Hey,
I see those things that need (at least) to be done/fixed for the next
Wed release:
- pinger does not respond to deletions (HAWKULAR-235/-263)
- inventory does not delete metrics on resource deletion HWKINVENT-45
(if INV can not make it, can UI send the delete metrics requests ?)
- duplicate URL addition HAWKULAR-255
- alerts is not multi-tenant yet (PR is sent, but may be too disruptive
?) HWKALERTS-40
-- Add url fails across tenants for same url (HAWKULAR-277 contains a
possible solution)
-- marking alerts as resolved (needs above PR for the backend)
- working pagination
- working availability charts
- move everything on non-snapshot releases
- update / write documentation on hawkular.org
As we will most probably not have full app servers functionality,
we should disable that tab so that users are not confused by it
and re-enable and work on it after the release.
Similar for the embedded agent.
Can we please concentrate on the above (and what else is needed that is
not yet on the list) to get
the release out?
9 years, 3 months