Build 154


On-Premises Release Notes

For an upgrade to the new version please follow the Packaged Based Operations manual.

Cumulative Features from Release #151, #152, #153 and #154

New Kubernetes Support

Beginning with this release we are rolling out our new version of the Kubernetes support. See Kubernetes for details.


We now support OpenShift versions 3.10 and 3.11.

Single-Page Application (SPA) Insights for End-User Monitoring

This release enriches our website end-user monitoring capabilities with single-page application (SPA) insights. It's now possible to track page transitions inside SPAs by leveraging our page API. Along with this release comes new terminology to better model the SPA paradigm; please read up on our documentation which explains the new terminology in detail.

New Single-Page Application Metrics

The details view of an issue contains a new "View Built-in Event" or "View Custom Event" link, which directly navigates to its corresponding definition in the Events & Alerts settings. This makes it easier to understand why a particular issue has been created.

New Python Tornado Instrumentation

With the latest Python package release (1.11.0), we now automatically instrument your Tornado applications. Please see the Python documentation for details.

Python Tornado Trace Detail

CockroachDB sensor

Instana now automatically discovers and monitors CockroachDB, a distributed cluster SQL database. Read more about our CockroachDB support.

Jenkins sensor

Instana now automatically discovers and monitors Jenkins, the leading open-source automation server. Read more about our Jenkins support.

Microsoft BizTalk Monitoring

A new sensor has been released along with the respective backend capabilities to monitor BizTalk. This is a first release, focusing on BizTalk hosts and their processing pipeline.

Docker container info

With this release, it is possible to view complete Docker container info for each container. The info is similar to docker inspect command output.

Internal calls

Calls will be created from SDK intermediate spans and their endpoint type are marked as "Internal". This feature is useful if you need to add additional information such as custom tags to the trace by sending intermediate spans through the Trace SDK. The calls and the tags are searchable in the Analyze View but will not contribute to the dashboard KPIs.

SAML IdP-initiated login

From this release on we also support IdP-initiated login. This allows users to select Instana from their IdP-specific selection screen and log them in directly. This doesn't require any reconfiguration so existing SAML-configurations won't have to be updated to use this.


Design Improvements

With this release the design of the top-level navigation, time selector and infrastructure dashboards is overhauled. This is the first of many steps to achieve better user interface consistency and usability.

The top-level navigation is turned into an expandable sidebar to clear up confusing navigation patterns and to make room for new capabilities that are released in the next months. As part of this we also improved presentation of the configured time. The configured time previously required a lot of room and was not always easy to read. This is now a lot better. Lastly the infrastructure dashboards were visually changed to be similar to the application and website monitoring dashboards.

infrastructure map with the new top-level navigation

New visuals for the infrastructure dashboards


  • Maintenance windows become effective immediately (if they are active at the current point of time) when they are created (both through UI and API). Previously Maintenance windows would have to be scheduled at least 4 minutes in advance.
  • When configuring custom events for endpoints one can now define events independently for synthetic and regular endpoints by selecting corresponding metric KPIs. Also built-in event rules for synthetic and regular endpoints were separated.
  • The effective connection type is collected in our website end-user monitoring.
  • Analyze View: We added the IS_BLANK and NOT_BLANK operators for filtering key-value tags.
  • Elasticsearch endpoint names for uncommon API operations are not "undefined" anymore.
  • You can use the real Kubernetes service name instead of the container name to name a service, but you will have to opt-in. Customer Success will be glad to help you.
  • Azure sensors can be disabled. Azure services discovery filtering by tags and resource groups can be performed. It is possible to define global filters for all Azure services but it is also possible to override global filters on sensor level.
  • Azure SQL replica databases can be filtered during service discovery.
  • Go Sensor: Simplified HTTP client tracing: New Helper HTTP HandleFunc: docs
  • Python Sensor: Asyncio: Optionally carry context along with Tasks; Configurable
  • Python Sensor: Stability and determinism improvements
  • Application Configuration: Applications configured with expression agent.tag equals something is redefined as agent.tag.something is present, the semantics remain the same.
  • Issue Severity: The severity of an issue remains at the maximum level that was reached during its lifetime and is therefore never decreased
  • Disk Capacity Detector: The issue of a disk with critical remaining capacity stays open even though the disk stops filling up or is out of capacity
  • If the parent span of an entry span is not present while a trace is processed, a call will still be created from the entry span instead of being discarded. For example, an SDK entry span is child of a jdbc exit span, when both spans are present when the trace is processed, one call will be created out of the two spans. Otherwise, one call will be created from the entry span and eventually another call will be created from the exit span.
  • The website monitoring JavaScript agent will now batch beacon transmissions for a short amount of time. This is done to more efficiently transfer data when HTTP/2 or keep-alive isn't possible.
  • The JavaScript agent's snippet now protects against accidentally duplicated inclusions on websites.
  • We will now apply SameSite=Lax cookie restrictions for increased security.
  • HAProxy sensor: Improve stability when encountering a large number of metrics.
  • WebLogic sensor: Improve stability when encountering a large number of metrics.
  • Built-in events: We improved the algorithms of built-in sudden drop/increase events in case of sparse input data.
  • Maintenance windows: Improve user experience by explicitly mentioning the timezone depending on the user's configuration.
  • Go Sensor: Update tests, docs and examples to use new repo name: go-sensor
  • Python Sensor: Improve JSON spans: Unify common elements
  • Ruby Sensor: Expand Legacy support for ffi, oj and sys-proctable
  • Ruby Sensor: Pass OJ options in to dump/load calls instead of setting them as default
  • AWS Sensors: Applied exponential backoff strategy for metrics gathering


  • Application perspective: Fixed broken dependency map links between services due to presence of intermediate spans.
  • Analyze View: Fixed searching with key-value pair tags with a single value.
  • Python Sensor: aiohttp: Explicitly set scope to None when not tracing
  • Python Sensor: Asyncio & Asynqp Cleanup, Tests and Minor Improvements
  • Python Sensor: Asynqp: Update and use the TEXT_MAP propagator
  • Python Sensor: Handle Server-Timing headers manually
  • Python Sensor: Asynqp: Support routing_key from named arg
  • Python Sensor: Asynqp: Protect against short tuples
  • Solr Sensor: Fixed an issue with collecting metrics regarding number of requests and cache (Solr version 4 and later)
  • Rollup Calculation: The error rate for Applications, Services and Endpoints shown in event details do now consider the distribution in calls
  • Dynamic Focus Query: Support querying on StatsD custom metrics when creating a Custom Event
  • Chart legend colors are now in sync with the colors shown in the tooltip when filtering metrics
  • Nginx Sensor: Fixed parser for map directive
  • AWS Sensors: Added self-recovery for missing IAM profile
  • WebHook and Splunk alert channel: Added missing entityLabel field in the generated alerts.
  • "Unspecified" service: Fix broken links to endpoint dashboards.
  • Dynamic Focus: The types application and endpoint are now listed in the suggestions for entity.type and entity.selfType.
  • Dynamic Focus: The event.type:change query matches only the change events, excluding online and offline events.
  • The website monitoring JavaScript agent's XMLHttpRequest instrumentation was changed so that it is compatible with more third-party trackers.
  • Go Sensor: Remove a code branch that is never followed in the Go Tracing Adapter
  • Python Sensor: Update aiohttp client error reporting