The biggest feature is that the back-end storage supports M3DB. M3 is a time-series database open sourced by uber. It claims to process 6.6 billion monitoring indicators inside uber. It is very convenient for expansion and contraction, and disaster tolerance is also very good. It defaults to Nightingale. Compared with the rrd version used, it has smaller requirements for hard disk IO, but it stores raw data and will take up more hard disk space, which is worth a try. Other upgrade points are as follows:
front end
-
fix: fix IE11 compatibility issues, currently supports IE >= 11, Chrome >= 70
-
fix(mon): Fix the problem that the shielding strategy cannot select shielded nodes
-
fix(mon): Fix the problem that some log collection changes will cause the name to be cut
-
feat: Add shortcut links between various systems
-
feat(mon): The monitor disk supports caching the number of display columns set by the disk
-
style: updated a new logo image
-
style(mon): re-categorize the monitoring system menu
rear end
-
Enhanced security: increased password complexity, optimized cookie processing, etc.
-
Support M3DB as a storage backend, please check the wiki introduction on github for details
-
Fix the problem of the alarm engine and the number of condition strings
-
The ability to add custom fields to the host device
There is a changelog in the code root directory, which lists the changed content and affected modules.