Prometheus exporter for AWS CloudWatch - Discovers services through AWS tags, gets CloudWatch metrics data and provides them as Prometheus metrics with AWS tags as labels
Bugfixes and features
Features:
Bugs:
Docs:
Refactoring:
Dependencies
New contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.59.0...v0.60.0
Important news and breaking changes
This release brings a bunch of breaking changes:
roundingPeriod
for discovery jobs is deprecated, a warning will be logged at startup. This is being deprecated in favor of always using the metric period. The implementation for roundingPeriod
can result in inconsistent Start and EndTime between batches. This negates its intent to ensure Start and EndTimes align with the metric period for CloudWatch best practices. This has the potential to produce data which will look inaccurate when compared against CloudWatch itself driving a lot of confusion. See https://github.com/nerdswords/yet-another-cloudwatch-exporter/issues/1290 for further context.delay
at the metric level is deprecated, a warning will be logged at startup. This delay
configuration has existed for a long time but was never utilized. Deprecating it and eventually removing it was chosen to simplify the configuration. See https://github.com/nerdswords/yet-another-cloudwatch-exporter/issues/1290#issuecomment-1948904375 for further context.type
field and the keys of exportedTagsOnMetrics
must be the AWS namespace rather than the alias (the README contains an up-to-date list of namespaces). Aliases are not allowed anymore. An error will be thrown at startup in an invalid namespace or an alias is used.aws_es_esreporting_failed_request_sys_err_count
is aws_es_reporting_failed_request_sys_err_count
aws_es_esreporting_failed_request_user_err_count
is aws_es_reporting_failed_request_user_err_count
aws_es_esreporting_request_count
is aws_es_reporting_request_count
aws_es_esreporting_success_count
is aws_es_reporting_success_count
aws_kafka_kafka_app_logs_disk_used
is aws_kafka_app_logs_disk_used
aws_kafka_kafka_data_logs_disk_used
is aws_kafka_data_logs_disk_used
aws_rds_rdsto_aurora_postgre_sqlreplica_lag
is aws_rds_to_aurora_postgre_sqlreplica_lag
aws_glue_glue_.*
is aws_glue_.*
These breaking changes will allow making the configuration easier to understand and less error prone, and also to build better documentation around supported services.
Bugfixes and features
Features:
discovery.exportedTagsOnMetrics
: validate that keys match one of the job types defined by @cristiangrecoRefactoring:
Dependencies
New contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.58.0...v0.59.0
Bugfixes and features
Features:
Bugs:
/aws/sagemaker/TrainingJobs
by @tristanburgessRefactoring:
GetMetricsData
processing by @kgeckhartGetMetricData
: refactor QueryID generation and result mapping by @kgeckhartpromutil.BuildNamespaceInfoMetrics()
and promutil.BuildMetrics()
into promutil.BuildMetricName()
by @tristanburgessDependencies
New contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.57.1...v0.58.0
Important news and breaking changes
Bugfixes and features
Bugs:
Dependencies
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.57.0...v0.57.1
Important news and breaking changes
includeContextOnInfoMetrics
can be used to include contextual information (account_id, region, and customTags) on "info" metrics and cloudwatch metrics. This can be particularly useful when cloudwatch metrics might not be present or when using "info" metrics to understand where your resources exist.apigateway:GET
permissions for ApiGateway discovery jobs, as that API is not being used anymore.Bugfixes and features
Features:
includeContextOnInfoMetrics
by @kgeckhartBugs:
Refactoring:
Dependencies
New contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.56.0...v0.57.0
Important news and breaking changes
list-metrics-callback
and max-dimensions-associator
feature flags have been removed: their behaviour is now the new default.Bugfixes and features
Features:
list-metrics-callback
by @cristiangrecomax-dimensions-associator
by @cristiangrecoBugs:
Refactoring:
Dependencies
New contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.55.0...v0.56.0
Important news and breaking changes
customNamespace
, which were deprecated in v0.51.0
, are now un-deprecated due to customers' feedbackalways-return-info-metrics
: return info metrics even if there are no CloudWatch metrics for the resource. This is useful if you want to get a complete picture of your estate, for example if you have some resources which have not yet been used.Bugfixes and features
Features:
Bugs:
Docs:
Refactoring:
Dependencies
New Contributors
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.54.1...v0.55.0
Bugfixes and features
Bugs:
Dependencies
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.54.0...v0.54.1
Bugfixes and features
Features:
log.format
flag by @cristiangrecoBugs:
FilterId
label value selection for s3 dashboard by @andriikushchDocs:
Refactoring:
Dependencies
New dependecies:
Updates:
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.53.0...v0.54.0
Bugfixes and features
Services:
Bugs:
Dependencies
Full Changelog: https://github.com/nerdswords/yet-another-cloudwatch-exporter/compare/v0.52.0...v0.53.0