Train Series Release Notes


New Features

  • support cassandra cluster configuration with loadbalancing policy DCAwareRoundRobinPolicy. ‘local_data_center’ defalut value is ‘’.

  • Allow metric_dimensions filter to filter on multiple dimension values, on alarms count endpoint (IE metric_dimension=hostname:host1|host2). This brings the query parameters into parity between the alarms list and alarms count endpoints. This also restores functionality that was available on the deprecated java api.

  • Support cassandra connection timeout option which set timeout when creating a new connection.

  • Configuration option db_per_tenant added for InfluxDB to allow data points to be written to dedicated tenant database where the database_name prefixes the tenant ID, e.g. monasca_tenantid.

  • Merge monasca-log-api source code into the monasca-api and enable logs endpoints.

  • Introduce configuration options that allow to enable/disable metrics and logs endpoints.

  • Dimensions names and values can be scoped by a timerange which can make dimension related queries to large databases much faster because only the relevant shards are searched. Users that upgrade their Monasca Grafana Datasource plugin to version 1.3.0 will benefit from this feature.

Upgrade Notes

  • Configuration option legacy_kafka_client_enabled added to allow working with both legacy kafka-python and new Confluent Kafka client. Please set message format version for the Kafka brokers to to avoid performance issues until all consumers are upgraded.

  • Upgrade Storm to 1.2.2 from 1.1.3.

  • Upgrade Apache Kafka from version 1.0.1 to 2.0.1. Please consult official upgrading notes for complete information on upgrading from previous versions.

Security Issues

  • InfluxDB 1.7.6 fixes a security issue in which monasca-api leaks dimensions across projects.

Other Notes

  • Unified response when content-type is incorect. Changed response for log part from 415 to 400 (Bad request, message - Bad content type. Must be application/json).