EUM Uncaught Error Breakdown
Uncaught errors as traces are helpful, but sometimes you don’t care about a single issue. When many teams are involved or many errors occur, it is much more useful to get an overview of the situation. This is where the new uncaught error breakdown comes in. This breakdown is located in the EUM dashboard and it helps you understand:
- What types or errors are occurring, and how many are there?
- In which browsers do they occur?
- On what pages do they occur?
This overview helps you tackle uncaught errors in a structured way. Start with the most common errors and work your way to the uncommon ones. Need more information? Hit the traces button to learn more about the error, e.g. see the stack trace. Combine this with the ability to see everything that happened for a specific page load by clicking on the “All traces belonging to this page load”-button.
EUM Resource Overview
Third-party resources (e.g. scripts and images) are often the culprit for slow page loads. With our new resource overview, you will find it much easier to identify slow resource providers.
Broken down by second-level domain, this table shows you all the resource providers that are actively used by your website. It will show how many times resources were loaded from this resource provider and at what latency.
In the resource provider dashboard, this continues with insights into caching statistics based on real-user data. See changes in caching statistics over time and how this relates to loading times.
EUM Span Search Keywords
With a large number of active users, you will end up in a situation where the unfiltered traces list is simply noisy. Where we previously only supported the span.content keyword to search for specific spans, we are now starting to expand support for span searches. Specifically, we are starting with EUM specific search keywords.
Using the new keywords, it is now possible to search by:
- Browser and operating system
- Geo data like continent, country, latitude and longitude
- Error details like the error message or stacktrace
- Custom meta data
- Page name
- Resource URL and caching state
- Page loads with specific timings (e.g. slow SSL connection setups)
Office 365 Chat Integration
With this release we are adding an integration for the Office 365 Team chat. Just configure a Webhook URL for your room (https://docs.microsoft.com/en-us/outlook/actionable-messages/actionable-messages-via-connectors#sending-actionable-messages-via-office-365-connectors) and add it to to the integrations settings in Instana!
Docker Container Port Mappings
The Docker container sidebar now has more detailed port mappings listing the local to container port mappings
Newly Supported Technologies
- Support Java tracing for S3, ehcache, IBM MQ, Spring Scheduled, Akka Http 10, WebMethods http
- PHP tracing now supports 32bit PHP on Linux
Introduce Apache HBase Database Dashboard
- General stats
- Region-Server stats
- Store/Block health
- Support monitoring JVMs that are run as systemd process with private PrivateTmp=yes.
- ActiveMQ Sensor now tracks server role (master/slave)
- OpenLDAP is now supported within containers
- Httpd Sensor now supports AIX (as IBM httpd)
- Add support for GRPC client/server span types and their correlation
- Add option to install the agent as a service using the one liner
- Improved UI performance, reduced flickering when maps are very dynamic
- The time picker now closes automatically when a valid time range is applied
- Host and container view switcher will now remain open when switching between host and container view.
- Zoom controls in 3D map now no longer work linearly so that zooming via controls works as expected.
- Tooltips now work as expected for small charts.
- Latency charts are now easier to read for services which have a small number of calls.
- Deselection of the min data set now works for latency charts.
- The connection lost notification is now reliable.
- More keyboard controls for our tables (up, down and space bar to expand rows).
- Break down EUM TCP time into TCP and SSL time.
- The API is now rate limited to 5,000 calls per hour. How many calls are left and when new calls can be executed is available in the HTTP response via the X-RateLimit-* headers.
- Httpd Sensor now handles Apache config case insensitive and considers absolute includes outside the config root directory
- Postgresql Sensor no longer miscounts query counts
- Fixed Akka message passing tracing in Scala 2.12
- Memcached Sensor now correctly calculates hit rate
- Agent misinterpreted —bridge=none for Docker
- Java Tracing: Fixed rare case of async spans attached to the wrong trace
- Java Tracing: Fixed rare case of async vertex spans not attached to a trace
- .Net Tracing: Fixed not so rare case of async spans not attached to a trace
- Fixed OpenTracing correlation between Java and Go
- Fixed Camel and Spring Batch traces rendering
- Ruby: loginfo/addinfo should properly translate span.kind #76
- Ruby: Lower logging verbosity & levels #78
- EUM: Resources with $ characters in the URL can now be tracked.
- EUM: Timeouts are now used for beacon transmissions.
- Search for terms that include white spaces and wildcards is not supported yet (e.g. ‘entity.label:”Shop *Service”’ does not work). If you remove the whitespace the search works (e.g. ‘entity.label:”Shop*Service”’ results in a working query).
- Top process list by host and Redis slow logs are only available in live mode.
- EUM error breakdown may list errors, browsers and pages multiple times for errors that have occurred before 2017-06-23 (UTC).