Github user cestella commented on the issue:
https://github.com/apache/metron/pull/840
I want to pile on and give this my (non-binding since I contributed PRs
against this PR) +1. LGTM!
---
Github user justinleet commented on the issue:
https://github.com/apache/metron/pull/840
At this point, I'm +1 since @merrimanr ran up the e2e tests. A couple
people have put a fair amount of testing into this, and it seems like at this
point we're at parity and not finding more
Github user merrimanr commented on the issue:
https://github.com/apache/metron/pull/840
I ran this up in full dev again and verified the e2e tests now work similar
to how they do in master. I also manually tested several other areas including
the Alerts UI, Kibana and Swagger.
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Worked with @merrimanr to fix the e2e test issue and just submitted a fix.
This does not fix the full e2e test runs as this is being handled by 857,
however it brings this PR back to parity with
Github user merrimanr commented on the issue:
https://github.com/apache/metron/pull/840
I spun this up in full dev and spent all day testing it. From a functional
perspective, I can not find anything wrong with it. I ran through the test
plan in this PR and everything worked as
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
This is going to take some time to resolve, but everything else in this
branch is still unhindered.
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Looking at the additional deployment readme merge conflicts now
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Just a status update on this. We're currently waiting for 0.4.2 to roll out
before this gets committed. We definitely want more eyes and testing on this PR
considering its breadth and size. We do
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
Yeah, I'm trying to think of the right way to do it, without having to have
it be "officially maintained".
more like an informal /testing_stuff , with some descriptions, when it was
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
I don't think that's a bad idea. I definitely like having test scripts
associated with the PR's because we can explicitly see what was done and/or
recommended at the time the PR was being ushered
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
So, this has me thinking It is a shame to have these test materials (
and others like @JonZeolla creates ) embedded in these pr's.
Maybe we should have someplace to put them in
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Ok, here is the remaining bit of the test plan for verifying ES. Thanks
@cestella for pulling together some of the public test scripts around this.
# Test Script
Testing Kibana
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
For reference, I've added some PDF snapshots of the dashboards to the Jira
-
https://issues.apache.org/jira/secure/attachment/12899952/Metron-Dashboard%20-%20Kibana.pdf
-
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Ok, the Metron error dashboard is in now. I'll add some additional testing
instructions tomorrow, but this PR should be ready for some more vigorous
testing. The most easily accessible e2e place to
Github user mraliagha commented on the issue:
https://github.com/apache/metron/pull/840
Yes, I agree. It completely makes sense to minimize the scope and work on
stabilizing this version at this moment.
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
For reference, here is a list of some of the follow-on work we should
consider:
- Improvements to Kibana dashboard
- Add new timestamp field to parsers and index templates to take place
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
@mraliagha I do think we should consider revisiting the field name
conventions, but I'd push for that as a follow-on task. As discussed in other
points on this thread, e.g. going straight to ES
Github user mraliagha commented on the issue:
https://github.com/apache/metron/pull/840
Is this the best time to ask for changing field name convention to avoid
dot or colon? We are externally using Hive external tables on HDFS data, due to
Hive limitations we need to change the
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
I was looking at that today as well, and I think that should be a follow-on
considering how large the change is moving from 2.x to 5.x. The changes
I've implemented (and Casey and Nick) for
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
6.0 was released, should we consider going to that while we are at it?
---
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
The only issue with it, is if we did that in the old version, then this
would be a regression. Other than that, I'm just waiting to review the docs ;)
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Whew, that's great news!
Hahahahaha, that's the default colors from the Kibana widgets... I usually
refrain from bringing it up, but I'm partially colorblind (even though I used
to do a
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
@ottobackwards You should be good to try it now.
---
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
I got full_dev failure on connecting to metron web. Can you post when you
think full dev is ready and I'll try again then?
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
Ok, looks like 2 more minor issues.
1. params.py missing the kibana_server_host, so it's not making its way to
kibana.yml
2. The default index mapping template that ES creates when I index
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
Thanks @mmiklavc, those changes look good. Running up in full dev
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
@mraliagha @cestella agreed on timestamp.
---
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
Can you call the curator stuff out in your checklist above?
---
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
Ok, killing my vagrant up then, until your commit
---
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
@ottobackwards Oh, afa the data pruner, I believe we can use Curator now as
a built-in mechanism for pruning data. Steps will be added to the documentation.
Github user mmiklavc commented on the issue:
https://github.com/apache/metron/pull/840
@ottobackwards thanks for the review! The templates themselves changed, but
the deployment of those templates is the same. I discovered that with a host
change in Kibana, I broke the Ambari
Github user ottobackwards commented on the issue:
https://github.com/apache/metron/pull/840
Are there any required changes to the ansible roles? I thought we injected
the templates from there. Or has the indexing service injection landed?
---
Github user cestella commented on the issue:
https://github.com/apache/metron/pull/840
@mraliagha Yeah, absolutely. This task purposefully leaves out major
changes that we could do to improve things (e.g. moving back to `.` separated
fields instead of `:`). The attempt is to touch
Github user mraliagha commented on the issue:
https://github.com/apache/metron/pull/840
It might not be completely related to this PR, but since _timestamp is
removed in ES 5.x, can we add a specific time of indexing at indexing bolt to
capture time of indexing. It is useful for the
34 matches
Mail list logo