hjp
/
ltsdb
1
0
Fork 0
Code Issues 3 Pull Requests Projects Releases Wiki Activity
72 Commits 1 Branch 8 Tags 284 KiB
Commit Graph

72 Commits

Author SHA1 Message Date
Peter J. Holzer c79f7de78e Merge branch 'master' of git.hjp.at:hjp/ltsdb 2023-02-04 12:19:19 +01:00
Peter J. Holzer ad601984d1 Add some debug messages 2023-02-04 12:19:01 +01:00
Peter J. Holzer d3a439bb49 Stop if we run out of tickmark frequencies 2023-02-04 12:16:37 +01:00
Peter J. Holzer 62b22d7516 Record OS version 2023-02-03 13:27:25 +01:00
Peter J. Holzer cfc514a3eb Merge branch 'master' of git.hjp.at:hjp/ltsdb 2023-01-26 21:28:24 +01:00
Peter J. Holzer 5dbe0391dd Extend linear time scale 2023-01-26 21:27:21 +01:00
Peter J. Holzer 75ce29135c Increase obsolescence limit to 1 week 2023-01-26 21:26:25 +01:00
Peter J. Holzer 125dd443c1 Record reboot overdue time 2023-01-14 19:20:57 +01:00
Peter J. Holzer 22c790080a Sort by healthscore 2023-01-10 21:34:44 +01:00
Peter J. Holzer 7ab45d3e90 Display min and max values in timeseries 2023-01-10 21:33:47 +01:00
Peter J. Holzer e57e333942 Make record_meminfo executable 2023-01-07 23:40:14 +01:00
Peter J. Holzer 52a5f60389 Record meminfo 2023-01-07 13:45:03 +01:00
Peter J. Holzer ed3f7872e9 Expose node 2023-01-07 13:26:40 +01:00
Peter J. Holzer fdf782bc2f Rename ltsdb_report.send_report to ltsdb_record.record_observations 2023-01-07 13:08:44 +01:00
Peter J. Holzer a83ebfcc99 Add new API endpoint /record as a (preferred) alias for /report 2023-01-07 12:55:21 +01:00
Peter J. Holzer ca0ce798da Keep min and max values of dropped data points 2023-01-05 15:14:28 +01:00
Peter J. Holzer c5b8476a42 Rename reporting/recording scripts 
I think record_* is less ambiguous than report_*.
A close second would be collect_*, but record seems to me to be a bit
closer to what the scripts are doing.
 2023-01-03 18:33:58 +01:00
Peter J. Holzer 25381714a1 Use special linear scale for time values 2023-01-01 01:06:24 +01:00
Peter J. Holzer f62816a250 Fix linear yscale 2023-01-01 01:05:38 +01:00
Peter J. Holzer 4380210c34 Display min, max and last value in description 2022-12-31 17:55:25 +01:00
Peter J. Holzer 4b6e6c6eed Use special log scale for time values 
Seconds are a bit unintuitive for everything above 1 hour.
 2022-12-31 17:51:38 +01:00
Peter J. Holzer fedc4c66ac Implement log scale for y axis 2022-12-31 16:58:38 +01:00
Peter J. Holzer d6bdaa4128 Don't try to render timeseries which are too short 2022-12-30 21:32:48 +01:00
Peter J. Holzer 0595c19de9 Avoid division by 0 
Not quite sure how that could happen but it did.
 2022-12-30 21:31:50 +01:00
Peter J. Holzer f272b1ba95 Color data points in timeseries according to their criticality 2022-12-30 20:48:39 +01:00
Peter J. Holzer 5dec92d736 Add tickmarks for 10 day intervals 2022-12-30 20:29:39 +01:00
Peter J. Holzer 2b0a0f6d64 Increase spacing between horizontal ticks a bit 2022-12-30 20:19:41 +01:00
Peter J. Holzer 8d02bf2c6b Add description to timeseries 2022-12-30 20:14:00 +01:00
Peter J. Holzer 32b9f8eeed Allow multi mode for timeseries 
Temporary until I think of something better.
 2022-12-30 19:43:56 +01:00
Peter J. Holzer 55a86772e0 Plot timeseries 2022-12-30 18:59:58 +01:00
Peter J. Holzer 28a7bbd4f3 Skip timeseries without recent data 2022-12-29 13:40:47 +01:00
Peter J. Holzer 851d9b4b37 Add logging to process_queue 2022-12-27 11:44:03 +01:00
Peter J. Holzer 0fab037cbc Extend time unit to years 2022-12-27 11:10:06 +01:00
Peter J. Holzer 74c6a19302 Implement descending criticality thresholds 2022-12-27 10:54:19 +01:00
Peter J. Holzer eeb7db96b8 Cap time until full at 1 billion seconds 
JSON can't represent infinity (at least not portably) so we need to use
a finite value. I certainly won't be worried if a disk will fill up in
32 years.
 2022-12-27 10:50:41 +01:00
Peter J. Holzer 42ba485794 Add dashboard with gauges 2022-12-27 10:29:49 +01:00
Peter J. Holzer 2f400fccee Print SSLCertVerificationError if we get one 
This also adds an option --verbose, but it doesn't actually do anything.
The error is alway reported.
 2022-12-21 10:03:13 +01:00
Peter J. Holzer b10b62e77d Make expiry exponential 
Choosing the element to expire from a uniform random distribution tends
to expire data points much too early. We want to always keep the oldest
observation and have a roughly exponential distribution between the
newest and the oldest observation.
 2022-12-17 17:49:19 +01:00
Peter J. Holzer a802f2ee27 Process queue 2022-12-11 22:58:26 +01:00
Peter J. Holzer caeaa0bf73 Add changed timeseries to queue 2022-12-11 22:58:04 +01:00
Peter J. Holzer 1e034f1a7f Make logging configurable 2022-12-11 22:57:41 +01:00
Peter J. Holzer 658fbdf83c Filter graphs by dimension=value 2022-12-09 22:55:07 +01:00
Peter J. Holzer cf32c72b8f Think about scheduling measurements and processing them 2022-11-27 10:19:37 +01:00
Peter J. Holzer dd1b3056ac Measure HTTP(S) response code and time 2022-11-27 10:18:23 +01:00
Peter J. Holzer 4bf3508d26 Factor out send_report 2022-11-27 10:17:43 +01:00
Peter J. Holzer 803a56d6df Think about authentication 2022-11-27 10:16:26 +01:00
Peter J. Holzer 009cf38adb Format timestamps explicitely 2022-11-26 23:36:00 +01:00
Peter J. Holzer e23f3f8726 Report errors as "just expired" 2022-11-20 18:46:39 +01:00
Peter J. Holzer 81fa412202 Add simple visualization 2022-11-20 18:43:45 +01:00
Peter J. Holzer a79aac5cfd Downgrade to Python 3.7 for Debian 10 2022-09-27 12:30:45 +02:00