App Metrics Release Notes and Known Issues

Page last updated:

This topic contains release notes for App Metrics 2.0 and later.

v2.0.5

Release Date: October 9, 2020

Requires TAS 2.10+, 2.9.5+, 2.8.11+, 2.7.17+ or 2.6.21+.

New Features

This release has the following new features:

  • When indicator documents are loaded, a test message is now sent to the defined webhook endpoints.

Resolved Issues

This release has the following fixes:

  • Webhook endpoints are no longer required to be Slack URLs. They will be sent to the webhook endpoint the format of: `{“text”: “[] violated threshold\nApp: \nIndicator: \nSince: \nCurrent value: \nDetails: ”}`

Maintenance Updates

This release updates the following release:

  • Log Store release to v4.0.14

v2.0.4

Release Date: September 9, 2020

Requires TAS 2.10+, 2.9.5+, 2.8.11+, 2.7.17+ or 2.6.21+.

Resolved Issues

This release has the following fixes:

v2.0.3

Release Date: August 3, 2020

Requires TAS 2.10+, 2.9.5+, 2.8.11+, 2.7.17+ or 2.6.21+ due to a UAA client scope change.

This release made a change to the default dashboards instance queries to display by instance_id instead of by process_instance_id. This was done for easier correlation with the CF CLI. Metric Store versions 1.4.3 and prior do not parse instance_id as part of http metrics so an update to Metric Store will be required to view this information.

Additionally this release also made a change to display all labels returned from a promQL request. These additional labels are needed to accurately view Spring Boot Metrics. Please verify the VMware Tanzu Application Service for VMs (TAS for VMs) or Pivotal Application Service (PAS) tile metric registrar configuration in regard to label filtering when emitting Spring Boot Metrics.

Known Issues

Resolved Issues

This release has the following fixes:

Maintenance Updates

This release updates the following release:

  • Log Store release to v3.0.2

v2.0.2

Release Date: June 22, 2020

Known Issues

This release has the following issues:

Resolved Issues

This release has the following fixes:

Maintenance Updates

This release updates the following release:

  • Log Store release to v3.0.0

v2.0.1

Release Date: May 5, 2020

Known Issues

This release has the following issues:

Resolved Issues

This release has the following fixes:

  • Single Sign-On not allowing user login
  • App Metrics failing to start when app security groups are in use
  • View in PCF Metrics link missing in Apps Manager
  • Smoke Tests Fail on Environments with skip ssl verify set to true

Maintenance Updates

This release moves App Metrics back to metrics.sys.DOMAIN

v2.0.0

Release Date: April 3, 2020

New Features in App Metrics v2.0

App Metrics v2.0 includes the following new features:

  • Spring Boot Micrometer and tagged metrics support
  • CAPI v3 multi-process and rolling deployment support
  • Declarative custom observability codification using Indicator Protocol
  • Create persistent, sharable dashboard customizations using Indicator Protocol
  • All-new, improved time selection experience
  • Improved data ingestion performance for live metrics and logs observation
  • Fault-tolerant, scalable logs and metrics data storage
  • Increased metrics data retention
  • Increased logs data retention
  • Improved data ingestion performance

Deprecated Features in App Metrics v2.0

App Metrics v2.0 removed the following features:

  • Trace Explorer
  • App Event Monitoring and Alerting Removed
  • CF Staging Failed Event Removed

Known Issues

The following sections describe the known issues in App Metrics v2.0.0.

Single Sign-On not allowing user login

Users attempting to log into App Metrics via Single Sign-On will be redirected to the login page without actually being logged into App Metrics.

This is due to the UAA_URL environment variable in the appmetrics app being set to https://uaa.<system_domain> instead of https://login.<system_domain>.

App Metrics failing to start when app security groups are in use

App Metrics is unable to connect to its underlying postgres database and will hang until the app is killed by CF if app security groups are in use.

Creating a security group with the below configuration and assigning it to the app metrics org (system) and space (app-metrics-v2) should allow the appmetrics app to start. json [ { "destination": "0.0.0.0-255.255.255.255", "protocol": "all" } ]

Unable to view time selector for apps with long routes

Apps with long routes or multiple routes could make the time selector scroll off the page and be inaccessible.

Promql Explorer unclickable in Firefox

The promql explorer is not clickable due to a CSS problem in Firefox.

Timeframe log file download is 1 KB

When you try to download logs for a given timeframe, an HTTP timeout can result in a 1 KB log file.

The default HTTP timeout is 180 seconds.

To work around this issue, change the TIMEOUT_SECONDS environment variable for the appmetrics application to a value greater than 180.

The appmetrics application is in the system org and app-metrics-v2 space.

CF Admin users unable to view all apps in App Metrics

The App Metrics client requests the cloud_controller.admin_read_only scope.

The App Metrics client does not have admin level permissions at this time and will require an update to TAS to add this.

For the time being, please add cloud_controller.admin_read_only to your admin users.

Any users logged into to App Metrics will need to log out and log back in after the change is made in UAA.

Hostname Verification is not ignored when skip-ssl-validation is true

The App Metrics client was still verifying cert hostname when skip-ssl-validation was true

Non-admin users unable to create monitors

The monitors endpoint requires a user to have admin level permissions, specifically doppler.firehose.

For the time being, please use a user with the doppler.firehose scope to create monitors.

Any users logged into to App Metrics will need to log out and log back in after the change is made in UAA.

App Metrics removed the path that Apps Manager checks to determine if App Metrics is installed on TAS.

Log Store is unavailable for some apps

Log Store may experience this when a particular source id has millions of logs and the connection times out.

This required and update of the log store release packaged in App Metrics.

Log Store is reporting high resource usage

Log Store 3.0.2 fixes an issue with the hinted hand-off that was causing high resource usage (memory, CPU, and/or disc usage) depending on the customer environment and machine size.

This requires and update of the log store release packaged in App Metrics.

Smoke Tests Fail on Environments with skip ssl verify set to true

The smoke tests did not take the ssl verify configuration into account.

The smoke test errand can be turned off for affected environments.

The status of App Metrics can be checked by visiting https://metrics.sys.DOMAIN/integration-status

Compatibility with VMware Tanzu Application Service

App Metrics v2.0 requires the following versions of VMware Tanzu Application Service:

  • v2.6: v2.6.x and later
  • v2.7: all patches within minor
  • v2.8: all patches within minor

Compatibility with Operations Manager

App Metrics v2.0 requires the following versions of Operations Manager:

  • v2.6: v2.6.x and later
  • v2.7: all patches within minor
  • v2.8: all patches within minor

Additional Items to Note

Network Metrics

Please note that if the app you are monitoring is not configured to accept HTTP traffic, no data will be displayed on the HTTP Request Count, HTTP Request Latency, or HTTP Request Errors charts.

Promql source_id tag

PromQL should always have the source_id tag for non-admin users. The source_id tag is used to verify access to the requested application and without it non-admin users will not recieve data back from the metric store. App Metrics supports using a $sourceId parameter in PromQL which automatically refers to the sourceId of the current app. Example: cpu{source_id=“$sourceId”}

Custom Metric Process Labels

App Metrics v2.0 introduces a new app hierarchy to support multiple process types from CAPI v3. Processes that contain apporpriately tagged envelopes will be available for metric filtering, and will be labled accordingly within the charts.

There are some metrics, including custom and network metrics, that may not have proper process tagging available. In these cases, sourceId={'YOUR_APP_ID'} may be used.