Fix KSM deprecation warning on A5

[mfpierre]

What does this PR do?

Fix the use of a deprecation method only compatible with A6

Motivation

Make the KSM check work again on A5

Review checklist

• PR has a meaningful title or PR has the no-changelog label attached
• Feature or bugfix has tests
• Git history is clean
• If PR impacts documentation, docs team has been notified or an issue has been opened on the documentation repo

Additional Notes

Is it ok to call self.warning in the init function?

[codecov-io]

Codecov Report

Merging #2317 into master will decrease coverage by 3.21%.
The diff coverage is 50%.

@@            Coverage Diff             @@
##           master    #2317      +/-   ##
==========================================
- Coverage   85.83%   82.62%   -3.22%     
==========================================
  Files         188        4     -184     
  Lines       12676      374   -12302     
  Branches     1324       72    -1252     
==========================================
- Hits        10881      309   -10572     
+ Misses       1410       44    -1366     
+ Partials      385       21     -364

Flag	Coverage Δ
#activemq_xml	?
#agent_metrics	?
#apache	?
#btrfs	?
#cacti	?
#cassandra_nodetool	?
#ceph	?
#cisco_aci	?
#cockroachdb	?
#consul	?
#couch	?
#couchbase	?
#datadog_checks_base	?
#datadog_checks_dev	?
#directory	?
#disk	?
#dns_check	?
#ecs_fargate	?
#envoy	?
#etcd	?
#fluentd	?
#gearmand	?
#go_expvar	?
#gunicorn	?
#hdfs_datanode	?
#hdfs_namenode	?
#http_check	?
#istio	?
#kube_dns	?
#kube_proxy	?
#kubernetes_state	82.62% <50%> (?)

[truthbk]

Looks good to me 👍

[truthbk]

Calling instance methods like warning in the __init__ could be risky if the object weren't properly initialized... However at this stage, after the super(), this doesn't look too harmful to me. Also, the constructor is unlikely to fail after the warning.