Metric Store Release Notes

This topic describes the changes in this minor release of Metric Store.

v1.5.5

Release Date: August 16, 2022

• Updated golang to v1.17.8
• Updated routing-release to 0.232.0
• Fixed issue where Metric Store was unable to remove a shard marked as pending to remove while new points were being written to the shard.

v1.5.4

Release Date: November 19, 2021

Added support for TAS v2.12.

v1.5.3

Release Date: September 15, 2021

• Fixed issue where the charts were not loading when the doppler.firehose was not in scope.
• Updated InfluxDb to v1.8.5

v1.5.2

Release Date: February 10, 2021

Use Ops Manager as the CA for all certificates to facilitate easier certificate rotation on Metric Store nodes.

v1.5.1

Release Date: November 18, 2020

• Replication factor defaulted to 2.
  • This doubles previous storage requirements because metrics going forward are replicated across nodes.
  • Metrics stored before this upgrade continue to exist, but only on a single node. This causes inconsistent query results for non-replicated metrics, depending on which node is queried.
• Include and use route registration health check.
• The replication factor is configurable from Ops Manager
• The max concurrent queries is configurable from Ops Manager
• Updated tile internal certificates to include missing SANs. These certificates are regenerated automatically.

v1.4.5

Release Date: February 10, 2020

Use Ops Manager as the CA for all certificates to facilitate easier certificate rotation on Metric Store nodes.

v1.4.4

Release Date: September 29, 2020

• Fixed issue where Metric Store was intermittently dropping points between nodes
• Added instance_id label to http_* metrics
• Removed empty errand metric-store-smoke-tests

v1.4.3

Release Date: August 17, 2020

• This release fixes a memory leak that caused Metric Store to slowly run out of memory because of not cleaning up closed network connection go routines.
• Updated debug and metrics endpoints to be on different ports.
• Updated golang to v1.13.12 and Prometheus to v2.19.1.

Known Issues

• Some users have seen points dropped when metrics are transferred between nodes.

v1.4.2

Release Date: April 28, 2020

Resolves issue with http_total metric to include correct node_index label value.

Known Issues

Some users have seen points dropped when metrics are transferred between nodes.

v1.4.1

Release Date: April 27, 2020

• Updated the opsmanager_syslog property to true to allow operators to egress logs to a syslog destination.

Known Issues

This release has an issue in the http_total metric when the Metric Store is running on multiple nodes. Metric Store incorrectly tracks a label node_index value as 0 as opposed to the true node index. Resolved release v1.4.2.

v1.4.0

Release Date: March 31, 2020

• Replication now detects availability zones (AZ). For those upgrading from beta versions, this causes a loss in metrics.
• Updated our CA properties to help users rotate both their Ops Manager and CredHub stored certificates.
• Updated read sample limit to 20,000,000. This was previously set to 1,000,000.
• Removed PromScraper to inject Metric Store metrics into Firehose.
  • Scraping Metric Store Prometheus endpoint by means of mTLS.

Known Issues

This release has an issue in the http_total metric when the Metric Store is running on multiple nodes. Metric Store incorrectly tracks a label node_index value as 0 as opposed to the true node index. Resolved in v1.4.2.

v0.3.0 (Beta)

Release Date: November 19, 2019

• Introduced http_total metrics and http_duration_seconds_* histogram for request latency.
• Fixed metric_storage_duration_days to report 0 when there is no data.
• Fixed degraded query performance when there are more than 500,000 points.

v0.2.0 (Beta)

Release Date: July 9, 2019

Features

See Overview

Known Issues

This release has the following known issues.