This page provides specific instructions for upgrading to each currently supported version of StackState. The instructions detail any significant changes that may impact how StackState runs after upgrade, such as a change in memory requirements or configuration.
Review the instructions provided below before you upgrade!
​Node sizing requirements have been increased.
The old stackstate-server
pod has been replaced by a number of separate pods. Custom configuration in values.yaml
should be updated:
Configured email details in stackstate.components.server.config
should be moved to stackstate.components.viewHealth.config
.
Other custom configuration in stackstate.components.server.config
should be moved to stackstate.components.api.config
.
A new mandatory parameter stackstate.baseUrl
has been added. This is the public URL of StackState (how StackState is reachable from external machines) and is exposed via the UI script API.
The file values.yaml
should be updated to include the new stackstate.baseUrl
parameter. The old stackstate.receiver.baseUrl
parameter has been deprecated and will be removed in the next release, however, when no stackstate.baseUrl
is provided in StackState v4.2, the configured stackstate.receiver.baseUrl
will be used instead.
The following configuration must be manually added after upgrade:
etc/application_stackstate.conf
New mandatory parameter stackstate.web.baseUrl
. This is the public URL of StackState (how StackState is reachable from external machines) and is exposed via the UI script API. You can manually create a system environment variable called STACKSTATE_BASE_URL
or add the value manually as a string in the file application_stackstate.conf
.
The following configuration changes must be manually processed if you are using a customised version of a file:
etc/stackstate-receiver/application.conf
Renamed the namespace stackstate
. This is now stackstate.receiver
.
Renamed the parameter apiKey
. This is now named apiKeys
and should be a list in the format [${stackstate.receiver.key}, ${?EXTRA_API_KEY}]
.
processmanager.conf
Added new parameter processes.kafkaToElasticsearch.topology-events
.
processmanager/kafka-topics.conf`
Added new section kafka.topics.sts_topology_events
.
Go to the StackState v4.1 docs site.
There are several changes to the processmanager.conf
file that must be manually processed if you are using a customised version of this file:
The sts-healthcheckuri
has been moved from port 7071 to 7080
The startup-check
block has been removed completely
Go to the StackState v4.0 docs site.
With this version the minimal system requirements for the StackState node of the production setup raised from 16GB to 20GB
The configuration processmanager-properties.conf
was merged into processmanager.conf
for both StackState and StackGraph. If you have changes to either one of those configuration files, you changes will need to be reapplied after upgrade.
For trace processing StackState Agent needs an upgrade to version 2.5.0.
This release deprecates the withCauseOf
topology query filter, in favor of the `Root
Cause Analysis` topology visualization setting. Stored views
which require make use of the withCauseOf
construct will need to be manually adapted.
New versions of StackPacks already contain these changes, for custom views, the following
script can be used in the StackState Analytics panel to list the views that need
migrating.
Graph.query { it.V().hasLabel("QueryView").forceLoadBarrier().filter(__.has("query", TextP.containing('withCauseOf'))).properties("name").value() }
In this release a new way of scripting propagation functions has been introduced so that the script APIs can be used. Propagation functions using the old script style will still work, but have been made read-only via the UI. Old style propagation functions can still be created via StackPacks, the CLI and API.