Connection module for Amazon CloudWatch
This module accepts explicit credentials but can also utilize IAM roles assigned to the instance trough Instance Profiles. Dynamic credentials are then automatically obtained from AWS API and no further configuration is necessary. More Information available at:
If IAM roles are not used you need to specify them either in a pillar or in the minion's config file:
cloudwatch.keyid: GKTADJGHEIQSXMKKRBJ08H cloudwatch.key: askdjghsdfjkghWupUjasdflkdfklgjsdfjajkghs
A region may also be specified in the configuration:
If a region is not specified, the default is us-east-1.
It's also possible to specify key, keyid and region via a profile, either as a passed in dict, or as a string to pull from pillars or minion config:
myprofile: keyid: GKTADJGHEIQSXMKKRBJ08H key: askdjghsdfjkghWupUjasdflkdfklgjsdfjajkghs region: us-east-1
convert_to_arn(arns, region=None, key=None, keyid=None, profile=None)¶
Convert a list of strings into actual arns. Converts convenience names such as 'scaling_policy:...'
salt '*' convert_to_arn 'scaling_policy:'
create_or_update_alarm(connection=None, name=None, metric=None, namespace=None, statistic=None, comparison=None, threshold=None, period=None, evaluation_periods=None, unit=None, description='', dimensions=None, alarm_actions=None, insufficient_data_actions=None, ok_actions=None, region=None, key=None, keyid=None, profile=None)¶
Create or update a cloudwatch alarm.
Dimensions must be a dict. If the value of Dimensions is a string, it will be json decoded to produce a dict. alarm_actions, insufficient_data_actions, and ok_actions must be lists of string. If the passed-in value is a string, it will be split on "," to produce a list. The strings themselves for alarm_actions, insufficient_data_actions, and ok_actions must be Amazon resource names (ARN's); however, this method also supports an arn lookup notation, as follows:
arn:aws:.... ARN as per http://docs.aws.amazon.com/general/latest/gr/aws-arns-and-namespaces.html scaling_policy:<as_name>:<scaling_policy_name> The named autoscale group scaling policy, for the named group (e.g. scaling_policy:my-asg:ScaleDown)
This is convenient for setting up autoscaling as follows. First specify a boto_asg.present state for an ASG with scaling_policies, and then set up boto_cloudwatch_alarm.present states which have alarm_actions that reference the scaling_policy.
salt myminion boto_cloudwatch.create_alarm name=myalarm ... region=us-east-1
delete_alarm(name, region=None, key=None, keyid=None, profile=None)¶
Delete a cloudwatch alarm
CLI example to delete a queue:
salt myminion boto_cloudwatch.delete_alarm myalarm region=us-east-1
get_alarm(name, region=None, key=None, keyid=None, profile=None)¶
Get alarm details. Also can be used to check to see if an alarm exists.
salt myminion boto_cloudwatch.get_alarm myalarm region=us-east-1
get_all_alarms(region=None, prefix=None, key=None, keyid=None, profile=None)¶
Get all alarm details. Produces results that can be used to create an sls file.
If prefix parameter is given, alarm names in the output will be prepended with the prefix; alarms that have the prefix will be skipped. This can be used to convert existing alarms to be managed by salt, as follows:
- Make a "backup" of all existing alarms
$ salt-call boto_cloudwatch.get_all_alarms --out=txt | sed "s/local: //" > legacy_alarms.sls
- Get all alarms with new prefixed names
$ salt-call boto_cloudwatch.get_all_alarms "prefix=**MANAGED BY SALT** " --out=txt | sed "s/local: //" > managed_alarms.sls
- Insert the managed alarms into cloudwatch
$ salt-call state.template managed_alarms.sls
Manually verify that the new alarms look right
Delete the original alarms $ sed s/present/absent/ legacy_alarms.sls > remove_legacy_alarms.sls $ salt-call state.template remove_legacy_alarms.sls
Get all alarms again, verify no changes $ salt-call boto_cloudwatch.get_all_alarms --out=txt | sed "s/local: //" > final_alarms.sls $ diff final_alarms.sls managed_alarms.sls
salt myminion boto_cloudwatch.get_all_alarms region=us-east-1 --out=txt