Releases: pingcap/tidb-insight
Basic Displaying
Implemented module explorer
, which provides several commands that can display basic information of collected data.
Multiprocessing on export/import
Use multiprocessing on Prometheus data export & import, reduce time usage on large clusters.
Support TiDB Server Info
Add support of getting TiDB server info, this is not working on v2.0.x or earlier versions, and will generate empty result.
Prometheus export/import
Added feature to export Prometheus metric data and re-import them on another instance, for detailed investigation.
Fix Integration w/ Ansible
Fixed various issues when integrating with tidb-ansible
.
Binary Ready
It's now ready to integrate to tidb-ansible
with binary release.
Ansible Integration (Initial features)
Add ability to integrate with tidb-ansible
.
- collect log files of TiDB modles
- collect config files of TiDB modules
- collect
perf
data of process
Fix unhandled exceptions since v0.1-test
pdctl: handle exceptions when PD is not running on the host (#8) * pdctl: handle exceptions when PD is not running on the host * fix duplicated function call
Basic functionality of collecting system info
This is the 1st working release of TiDB-Insight, which now supports collecting:
- Basic system information
- Partition information & stats
- Progress info of TiPD/TiDB/TiKV server (multiple instance on one host is not supported yet)
- Runtime trace data using
perf
(disabled by default) - Application logs of TiPD/TiDB/TiKV server and system log (disabled by default)
- Dump critical system configurations (disabled by default)
- Runtime information of cluster from PD's API (enabled by default, optionally using
--pd-host
and--pd-port
if there's nopd-server
running on localhost)
Collected data are saved to data
under the project root.
To build the collector
binary, install Go and run make
.