November 6, 2024
Features
Merge natural language search with existing search bar
Earlier we had two different searches where users can type in natural language and users can search by selecting attributes , we merged them into one.
How it works:
If the user types anything and it doesn’t match with listed attributes then it will work as natural language search , if typed text is matched with attributes then attributes will popup as dropdown in suggestion and it will work as normal search.
This behavior is in infra , logs , apm search where we had natural language search. For other modules it will work as normal search only.
Azure integration (live on 7-11)
We now have Azure integration for Azure VM Scale Set, Container Service and Load balancer services.
Changes
- Updated JSON viewer component inside APM.
- Added ability to drag size of span name in waterfall view.
- New UI for flamegraph with better visibility , more options is live.
- Fixed the filter in live mode for logs. Now new data will come in live mode as per applied filters.
- Fixed the column resize issue while we change column size in the table in APM.
- Fixed the issue of span names not displaying correctly.
- Fixed the issue of alert not working in synthetic test.
- Node.JS APM : Now we are collecting console logs and it will be automatically correlated with traces.
LLM observability
Now we are published under integration section of traceloop and openlit so their users can directly view instruction to integrate Middleware.
API and Ingestion keys
Now we display ingestion keys in settings (which is useful for data ingestion via API or to access MW’s APIs)
- Optimized integration code (for MySQL, MongoDB etc) and added test cases.
- APM: added parent span ID in trace drill down.
Alerts
- Improved drill down of log alert by displaying relevant logs with filter (applied) as per alert rule.
- Improved alert email information by displaying log samples as with filters as per alert rule.
Need assistance or want to learn more about Middleware? Contact our support team in Slack.