We have open source tech stack in our firm. We use fluentd as log collector. How can I visualize my fluentd logs? Which are good open source tools for log analysis and visualization. I have researched many but none seems to fit in. I am not interested in free trial versions of enterprise offerings. I want something under open source license. Any help will be appreciated.
As far as I know you could use:
Fluentd + ElasticSearch + Kibana
Fluentd + Prometheus
Fluentd has plugin for both ElasticSearch and Prometheus (https://docs.fluentd.org/v0.12/articles/monitoring-prometheus)
Related
I can see that I can create/upload parquet files on on premise ssis with Azure Feature pack to my azure storage account.
However there is a lack of settings for parquet files, like partition or compression settings. The last one only available for text files.
what are my options to put a snappy compressed parquet file on my storage without databricks, just from my on premise ssis to storage account.
I tried it and you are correct, this is not supported yet.
You can share your feedback here so that the product group can look into this.
Alternately, here are some helpful docs you can pursue. Or perform compression on files in SSIS using the ZipFile Class in the recent version of the .NET framework.
I can't find clear information about the differences between Ejabberd Business and Ejabberd community (apart from the list of RFC/XEP here : https://www.process-one.net/en/ejabberd/protocols/).
Especially about the rest API : does community edition support it ?
In the documentation (https://docs.ejabberd.im/developer/ejabberd-api/) there is no info about which functionnality is supported by which version.
I also found this page : https://feedback.process-one.net/support/solutions/folders/6000076557 on the Business edition features but it seems to be outdated.
Thank you for your help.
Regards
If you enable mod_http_api, it provides ReST API in JSON format. Alternatively, if you install mod_rest from ejabberd-contrib git repository, you can execute commands with HTTP POST queries.
Is it possible to clear the parquet metadata cache in Apache Drill?
I'm finding that when you version files and use maxdir(), it holds on to old versions which may be deleted.
It seems like there should be a "clear metadata cache" command to cancel out the "refresh metadata cache" command.
Unfortunately, there is no such functionality in Drill so far.
Please log Jira for improvement [1]. Contributions always welcome.
[1] https://issues.apache.org/jira/projects/DRILL
my boss has asked me to find a tool that will monitor our sever health. Some kind of desktop application preferably that we can keep an eye on and will monitor us when capacity goes over a certain level, or we approach max storage etc.
We need to monitor both MySQL and Apache. I'm guessing I might need two tools.
THanks in advance
Have you looked at munin? it's not desktop... but i don't know why do you want to have a desktop solution?
you can monitor apache with SNMP module like mod_apache_snmp and tools like OpenNMS, and Nagios. Nagios supports monitoring mysql also.
You might also like to look at Megamon (http://www.megamon.com). Megamon is a complete monitoring solution capable of graphing numerous system performance aspects as well as escalating alerts and much more.
Megamon is not a desktop solution but runs as a Virtual Appliance. However, since it has a web interface, it is just as easy to use as a desktop application.
Have a look at SeaLion. SeaLion is a cloud based Linux server monitoring tool. Getting started is as easy as executing a command. It installs an agent at /usr/local/sealion-agent and runs as an unprivileged user (sealion). This agent will collect data at regular intervals across servers and this data will be available on your workspace. The latest version is shipped with NGINX, Apache, MySQL, MongoDB and Redis monitoring capability. It is free for 1 server with a 12 hours data retention policy.
I have an application running on Apache 2.2, tomcat 6 and it uses struts framework, hibernate framework. We use Mysql at back end. We also interact with third party servers to place some requests that are requested by the user. Due to confidentiality constrains I can not tell you exactly what we do but I can assure you that we have not customized any thing and we use the most general builds availaible for Apache, tomcat etc. We use Linux platform. Lots of visitors visit our site, where they first pay using payment gateways, and then buy a product. To buy the product we again have to hit a third party site. Its a simple e-commerce kind of setup. The problem is that some times the server hangs. As in it does not responds and when we click on a link that (I know) is served by the tomcat container, it does not get loaded. Here is what i need help for:
Since my hosting is on a headless linux platform, please suggest me a good debugging tool.
We have logging in place and we print stack trace of almost all exception(if they happen), we always monitor catalina.out, but when the server hangs, we dont see any activity on catalina.out. may be this can give some one a clue.
We have show_sql disabled for hibernate, we tried to enable it but still that was not sufficient to figure out if the application stucks on a query. We also have slow_sql log enabled but that does not show any significant queries. How can we check if my application is stuck on a query?
If my application is not stuck on a query, how can i know where it is stuck?
How can I get java stack dump?
What are possible ways to resolve such a problem.
ANy suggestions are welcome. I thank you all in advance for reading my question and for the time you will devote writing an answer.
Facing your situation I would do a thread dump when the servlet container "hangs". A thread dump provides you a list of the stack traces for all Java threads in a given JVM.
Doing a thread dump is pretty easy:
Find the Tomcat's process ID, e.g. ps -ef | grep java
Send a SIGQUIT to the process, e.g. kill -3 tomcatProcessId
You will find the thread dump's content in TOMCAT_HOME/logs/catalina.out.