View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0005487||Taler||deployment and operations||public||2018-11-27 10:53||2020-07-05 00:07|
|Reporter||Florian Dold||Assigned To|
|Priority||normal||Severity||feature||Reproducibility||have not tried|
|Target Version||Fixed in Version|
|Summary||0005487: set up system monitoring on gv|
|Description||On tripwire, we previously ran into the issue of a full /tmp directory.|
With gv, we should have monitoring in place to inform us of abnormal states, such as full disks or excessive memory use.
|Tags||No tags attached.|
||ack, will get to this after GSoC.|
||Please list all things to monitor here. They do not have to be in one note, but as you think of them, add them so we have a full picture when it's time to start this.|
Please create a list of things we should monitor. The list will help determine the best solution. If it's just RAM and disk space, we could have a simple BASH script and cron job. If the list is larger, we do something else.
Also Christian should be involved because I think he will probably use it (as root) most often. Maybe he likes something that he doesn't have to check, but notifies by e-mail when there is a problem. Maybe he prefers a GUI like Zabbix. Etc. But let's start with a list of things to monitor.
We need two things:
1. General system monitoring, such as disks, memory, CPU usage.
2. Application-specific monitoring, which will go over log files and extract "stuff" from it. What this stuff is depends to some degree on the application, but we want to at least collect log lines that are at the ERROR log level and alert if they occur with a too high frequency.
NetData (https://github.com/netdata/netdata) might be a good start for system monitoring.
Ah. Well if you already have a project in mind that you want to use, it's now a matter of how Christian feels.
This is the quickstart: https://github.com/netdata/netdata#quickstart
For the .deb: https://github.com/netdata/netdata/blob/master/packaging/installer/README.md
And a good overview: https://github.com/netdata/netdata/blob/master/docs/guides/step-by-step/step-00.md#netdata-fundamentals
Here is information on opting out of the analytics: https://github.com/netdata/netdata/blob/master/docs/anonymous-statistics.md
From the website it looks to have some nice features, including integration with PostgreSQL, nginx, syslog.
But I really can't imagine doing this on a live system without testing on a stage server first. But the next step is gettin Christian's opinion. Assigning to him.
||netdata seems fine. I do see the point that we need to sort out server staging. Don't have a good plan for that yet.|
> do see the point that we need to sort out server staging. Don't have a good plan for that yet.
I recommend a cheap VPS. They can be *very* cheap. We don't need anything fast, or even very secure just to do staging (except deployment.git stuff maybe). All we need to do is to duplicate the taler.net environment reasonably well.
|2018-11-27 10:53||Florian Dold||New Issue|
|2019-07-15 18:55||nikita||Assigned To||=> nikita|
|2019-07-15 18:55||nikita||Status||new => assigned|
|2019-07-15 18:55||nikita||Note Added: 0014680|
|2020-04-13 02:45||Christian Grothoff||Assigned To||nikita => buckE|
|2020-04-13 02:45||Christian Grothoff||Severity||minor => feature|
|2020-04-13 03:37||buckE||Note Added: 0015634|
|2020-06-19 08:54||buckE||Note Added: 0016320|
|2020-06-19 08:54||buckE||Assigned To||buckE => Florian Dold|
|2020-06-24 18:18||Florian Dold||Note Added: 0016356|
|2020-06-25 07:52||buckE||Note Added: 0016358|
|2020-06-25 07:52||buckE||Assigned To||Florian Dold => Christian Grothoff|
|2020-06-25 09:16||Christian Grothoff||Note Added: 0016361|
|2020-06-26 11:27||buckE||Note Added: 0016369|
|2020-07-05 00:07||Christian Grothoff||Assigned To||Christian Grothoff =>|
|2020-07-05 00:07||Christian Grothoff||Status||assigned => confirmed|