Change check

This section explains the change checklist of each module

Preface

There are two ways to determine whether the change is abnormal:

  1. Determine whether the process after the change is still alive. This method can be achieved by adding process monitoring to the module;
  2. By observing whether the indicators of the module are normal. This section focuses on the monitoring indicators that each module needs to pay attention to.

After each component of Nightingale is started, it will report some of its own status data to the monitoring system. You can observe whether the system is operating normally through the reported indicators. The monitoring indicators of each module are introduced below.

transfer change

The core indicators of transfer are as follows, with a statistical period of 10s.

Monitoring indicators meaning
n9e.transfer.points.in Number of receiving points
n9e.transfer.points.out.tsdb Number of points sent to tsdb
n9e.transfer.points.out.tsdb.err Number of failed points sent to tsdb
n9e.transfer.points.out.judge Number of points sent to judge
n9e.transfer.points.out.judge.err Number of failed points sent to judge
n9e.transfer.stra.count Get the number of monitoring strategies

If there is no obvious change in the above indexes of the transfer cluster after the change of transfer, then the change is in line with expectations.

tsdb change

The core indicators of tsdb are as follows, with a statistical period of 10s

Monitoring indicators meaning
n9e.tsdb.points.in Number of receiving points
n9e.tsdb.query.miss Number of times the query data is empty
n9e.tsdb.index.out.err Number of failed index pushes

If there is no obvious change in the above indicators of the tsdb cluster after the change of tsdb, then the change is in line with expectations.

index change

The core indicators of index are as follows, with a statistical period of 10s

Monitoring indicators meaning
n9e.index.query.counter.miss Number of misses when querying index with fullmatch
n9e.index.xclude.miss xclude query index misses

If there is no obvious change in the index data of the index cluster after the index is changed, the change is in line with expectations.

judge change

The core indicators of judge are as follows, the statistical period is 10s

Monitoring indicators meaning
n9e.judge.push.in Number of receiving points
n9e.judge.running Number of judge tasks being executed
n9e.judge.stra.count Number of strategies acquired

If there is no obvious change in the above indicators of the judge cluster after the change of judge, it means that the change is in line with expectations.

Last modified April 12, 2020: trans content/en/docs/operation (#7) (7f039b5)