Skip to content

Latest commit

 

History

History
70 lines (51 loc) · 4.82 KB

LegacyConfiguration.md

File metadata and controls

70 lines (51 loc) · 4.82 KB
Raw

Legacy configuration

DynamoDb

Field Type Value Description
Threshold Decimal Default is 0.8 (ie 80%) The used as a percentage of the capacity to work out the alarm thresholds. Will be used as threshold of all tables listed in an AlertingGroup unless overridden for a table.
MonitorThrottling bool Default is false Enables monitoring on throttled reads or writes.
ThrottlingThreshold int Optional Set the throttling threshold - the number of throttled reads or writes in a minute that causes an alarm.
Tables Array of Table ** Array of dynamo tables to add the alerts to.
See below for more info on Tables
ExcludeTablesPrefixedWith Array of strings Don't add read or write alerts for any tables with these name prefixes.
Exclude overrides Tables settings.
ExcludeReadsForTablesPrefixedWith Array of strings Don't add read alerts for any tables with these name prefixes.
Exclude overrides Tables settings.)
ExcludeWritesForTablesPrefixedWith Array of strings Don't add write alerts for any tables with these name prefixes.
Exclude overrides Tables settings.

Tables

Tables can either be added as strings or Table objects. If added as simple strings, the default options will used.

Field Type Value Description
Name String One of Name or Pattern must be specified Name of the dynamo table
Pattern String One of Name or Pattern must be specified Regex to match multiple dynamo tables
Threshold Decimal Default from the Threshold value of containing alerting group. Used as a fraction of the capacity to work out the alarm threshold for this table ***
MonitorWrites Boolean Default is true If false, no alerts are generated for writes to the table or its indexes
MonitorThrottling bool Default from the MonitorThrottling value of containing alerting group. Enables or disables monitoring on throttled reads or writes.
ThrottlingThreshold int Default from the ThrottlingThreshold value of containing alerting group. Set the throttling threshold - the number of throttled reads or writes in a minute that causes an alarm.

*** The alerting Threshold is a value from 0.0 to 1.0 that specifies the fraction of the table's read or write capacity being used that triggers the cloudwatch alarm. The threshold for a table uses these fallbacks:

  • if a threshold is specified for the table, that is used. If not,
  • if a threshold is specified for the containing alerting group, that is used. If not,
  • a global default of 0.8 (i.e. 80% utilisation) is used.

All global secondary indexes on the table will also be monitored at the same threshold.

Sqs configuration sections

Sqs

Field Type Value Description
LengthThreshold int, count of messages Optional Value for alert on queue length
OldestMessageThreshold int, number of seconds Optional Value for alert on old messages
Errors error queue data Optional Default configuration of error queues
Queues Array of queue

Queue

Field Type Value Description
Name String One of Name or Pattern must be specified Name of the queue
Pattern String One of Name or Pattern must be specified Regex to match multiple queues
LengthThreshold int, count of messages Optional Value for alert on queue length
OldestMessageThreshold int, number of seconds Optional Value for alert on age of messages in a queue
Errors error queue data Optional Values for alerts on error queues

Errors

Can be attached to Queue or alerting group or both. Values are defaulted - if a value is not specified in a queue, the default from the alerting group is used. If that is not specified, then a global default is used.

Field Type Value Description
Monitored bool Optional, default true Should this error queue be monitored at all. If False, none of the other values in the Errors object are used
Suffix string Optional, default is "_error" Text at the end of queue name to determine if this is an error queue
LengthThreshold int, count of messages Optional Value for alert on error queue length
OldestMessageThreshold int, number of seconds Optional Value for alert on age of messages in an error queue

This looks complex, but the defaults are overrides are aimed at producing readable and sensible markup, while dealing with the way that queues are set up in a variety of ways: in some cases the team needs to know about even 1 message arriving on an error queue, while in others the error queue is ignored, or does not exist.