mirror of
https://github.com/etcd-io/etcd.git
synced 2024-09-27 06:25:44 +00:00
Documentation/etcd-mixin/mixin.libsonnet: Add alerts for etcd fsync duration
This commit adds support to check the 99th percentile of the etcd members fsync duration and fires a critical alert when it is greater than 1 sec. The recommended fsync for etcd is 20 ms but there might be scenarios where a user might be using bad disks for reasons. This will make sure to let the user/admin know that it is critical for etcd performance.
This commit is contained in:
parent
73e5714bc5
commit
3022bd73ce
@ -173,6 +173,20 @@
|
||||
summary: 'etcd cluster 99th percentile fsync durations are too high.',
|
||||
},
|
||||
},
|
||||
{
|
||||
alert: 'etcdHighFsyncDurations',
|
||||
expr: |||
|
||||
histogram_quantile(0.99, rate(etcd_disk_wal_fsync_duration_seconds_bucket{%(etcd_selector)s}[5m]))
|
||||
> 1
|
||||
||| % $._config,
|
||||
'for': '10m',
|
||||
labels: {
|
||||
severity: 'critical',
|
||||
},
|
||||
annotations: {
|
||||
message: 'etcd cluster "{{ $labels.job }}": 99th percentile fync durations are {{ $value }}s on etcd instance {{ $labels.instance }}.',
|
||||
},
|
||||
},
|
||||
{
|
||||
alert: 'etcdHighCommitDurations',
|
||||
expr: |||
|
||||
|
Loading…
x
Reference in New Issue
Block a user